The collaboration with AI is not a recent development, yet the rise of foundation models marks a new chapter in the evolution of large language models (LLMs) capable of reasoning. Evidence is mounting that these advanced models can carry out clinical tasks with competence akin to that of human professionals.

India is progressing towards an AI-ready healthcare system, bolstered by governmental support, a vibrant start-up landscape, increased adoption of electronic medical records (EMRs), private investment, and a growing acceptance of AI technologies among clinicians.

Presently, the value generated from AI applications is largely focused on a limited number of use cases, such as workflow automation and ambient scribing. These initiatives aim to enhance efficiency and patient experience, although demonstrating a clear return on investment (ROI) for businesses in the short term proves challenging.

Despite noticeable demand, various promising AI applications still show significant potential due to the wide gaps existing in the market. Opportunities abound in areas like remote patient monitoring, optimization of operating rooms and intensive care units (ICUs), management of chronic diseases following hospital discharge, and claims processing—making these attractive prospects for AI-focused startups.

Nevertheless, the readiness to implement these technologies is inconsistent. Key obstacles such as data quality, effective workflow integration, clinician trust, and managing change hinder large-scale adoption. Successful start-ups typically possess a strong understanding of the healthcare sector combined with robust AI capabilities.

For those interested in further insights, a comprehensive report is available for download (PDF). 

HealthQuad, one of the foremost investment platforms in India focusing on innovative healthcare models, is actively investing from its HealthQuad Fund III. The organization continues the successful strategy and track record established in its previous funds. Founded in 2016, HealthQuad manages assets exceeding US$500 million and operates as the early-growth investment branch of Quadria Group, a prominent healthcare investment firm in Asia. HealthQuad supports outstanding founders who are creating category-defining companies in sectors such as HealthTech, MedTech, Bio/PharmaTech, and innovative care delivery, leveraging deep industry knowledge and operational capabilities to foster growth both in India and on a global scale.

Artificial Intelligence in Healthcare Services in India

A sunlit Indian clinic room where a doctor in a crisp white coat leans over a sleek tablet glowing with colorful neural-network overlays of a patient’s vitals. The adult patient sits calmly on the examination table as soft natural light streams through a nearby window, illuminating charts on the wall and a small Indian flag discreetly displayed. Subtle digital data pulses seem to flow from the tablet into the room, symbolizing AI’s life-saving reach in rural and urban healthcare alike.

AI in Indian Healthcare Delivery - Bain & Company

- The emergence of reasoning LLMs demonstrates human-level competence in clinical tasks.  
- India's AI-ready healthcare infrastructure is enhancing through government and start-up initiatives.  
- Barriers to AI implementation include data quality, clinician trust, and integration challenges. 

Written in collaboration with Written in collaboration with AI is not new, but foundation models are: The agentic era of reasoning large language models (LLMs) is emerging, with mounting evidence that clinical tasks can be performed at a human level of competence.





India’s healthcare infrastructure is becoming AI-ready, supported by government initiatives and a thriving start-up ecosystem, as well as rising electronic medical record (EMR) penetration, private capital, and clinician acceptance.





In AI deployments, value is currently concentrated in a narrow set of use cases (e.g., workflow automation and ambient scribing, among others) directed toward efficiency and patient experience gains while business ROI remains difficult to demonstrate in near-term.





Several high-potential AI use cases present large whitespace despite clear demand—including remote patient monitoring, operation theatre and intensive care unit (ICU) optimization, post-discharge chronic disease management, and claims management—creating compelling opportunities for AI start-ups.





Implementation readiness remains uneven—data quality, workflow integration, clinician trust, and change management remain the primary barriers to scale, favoring start-ups with deep healthcare domain expertise alongside AI capabilities. Download the report (PDF) About HealthQuad HealthQuad is one of India’s leading investment platforms focused on new-age healthcare models, and is currently investing from HealthQuad Fund III, continuing the strategy and track record established across Funds I and II. Founded in 2016, HealthQuad currently has an AUM of over US$500 million, and acts the early-growth investment arm of Quadria Group, one of Asia’s leading dedicated healthcare investment platforms. HealthQuad backs exceptional founders building category-defining companies across HealthTech, MedTech, Bio/PharmaTech and novel care delivery, leveraging deep sector expertise, operating capabilities and Quadria’s extensive healthcare ecosystem to scale in India and globally.

AI in Indian Healthcare Delivery

In the healthcare landscape of 2026, the conversation has shifted from the anticipation of AI's arrival to its rapid integration. As adoption increases, the focus on ensuring reliability and proper governance of these artificial intelligence systems is more crucial than ever. Understanding the capabilities of an AI model is vital, but equally important is the assurance that health systems can effectively implement and maintain these technologies. This stage is critical and hinges on the integrity of patient data, which serves as the backbone for successful AI deployment.

The safety of AI systems is directly tied to robust data quality management and governance. This gives rise to two fundamental questions regarding data quality in healthcare AI: First, is the data used for developing and validating AI models reliable? Second, is the real-time patient data the models engage with accurate, comprehensive, current, and associated with the correct individual? Relying on mere confidence is insufficient; in the next 12 to 18 months, healthcare providers can expect a closer integration of data quality initiatives with AI approval protocols, where substandard data may formally halt or delay AI rollouts.

According to a survey conducted by the American Medical Association in March 2026, 81% of physicians now utilize some form of AI in their practice, a significant increase from 38% in 2023. It’s important to clarify that this doesn’t mean that 81% are engaging in AI-assisted medical practice; rather, they are recognizing AI's utility in everyday clinical tasks. This includes activities such as summarizing research and treatment protocols (39%), generating discharge instructions and care plans (30%), and documenting billing codes and patient notes (28%).

The shift towards routine use of automation in administrative functions and workflows signals a significant trend. Nevertheless, the journey is far from complete. Innovations in generative AI are making inroads into electronic health records (EHR), while predictive AI is enhancing operational areas like risk assessment for re-admissions and identifying high-risk patients, along with scheduling and billing optimizations. Moreover, diagnostic AI, which encompasses computer vision and machine learning algorithms, has been integrated into medical device operations for some time.

For healthcare systems to be truly ready for AI integration, it is essential to rely on precise, accurate, and continuously updated data. The risks associated with flawed data quality can lead to catastrophic errors. For instance, if an AI system incorrectly prescribes a medication due to inaccurate dosage information, the consequences could be dire. Similarly, incorrect metadata regarding drug administration routes can result in administering potentially dangerous doses.

Other potential pitfalls include inaccurate risk assessments, misrouted health alerts, and false eligibility determinations. These issues can lead to financial losses, damage to trust in healthcare systems, delays in care, wrongful treatments, or even fatalities.

Duplicate patient records pose significant challenges for both healthcare research and clinical practice, with some analysts estimating a duplication rate of 8-10%. This means that a healthcare system with one million patient records could be grappling with 80,000 duplicates, complicating cost management and the accuracy of AI-driven decisions. Further complicating matters, fragmented records can elevate risks associated with AI applications, such as missing critical medications or contraindicated supplements.

Before an AI system can accurately interpret a patient's medical history, healthcare organizations must establish trust in the records associated with that patient. Techniques like identity resolution, record matching, and deduplication play a crucial role in creating a reliable, unified patient record. These practices not only enhance the quality of care but also serve to ensure that the data used in AI processes is solid.

Some of the most significant safeguards for AI technologies can be established well before the algorithms are applied. Solutions from companies like Melissa assist in achieving compliance with initiatives like Project US@. This project, overseen by the Office of the National Coordinator for Health Information Technology (ONC), standardizes patient address data to minimize inconsistencies in identifying information that can lead to fragmented or duplicated records.

Thus, data quality infrastructure must be treated as integral to AI infrastructure. An exceptional AI model could still yield inaccurate results if the data it processes does not accurately reflect patient realities. Consequently, issues such as duplicate records, wrong identities, and outdated medical information can have amplified repercussions in an AI context. 

Building a robust data foundation doesn’t require healthcare organizations to overhaul their existing data environments. A simple checklist can guide the process through all stages of data quality, starting from the beginning and extending to ongoing data oversight:

1. Validate and standardize identity attributes during patient registration, ensuring alignment with Project US@ standards.
2. Resolve and deduplicate data across EHR, claims, and ancillary systems prior to feeding that data into AI models.
3. Assess the consistency and currency of clinical data fields that AI utilizes, such as medications, allergies, and coverage information.
4. Maintain vigilant monitoring of identity discrepancies and address them as AI-related incidents, rather than relegating them to routine administration.
5. Include evidence of data integrity in the AI evaluation process alongside model performance metrics.

Readiness for AI involves more than treating data quality as an intermittent task. Healthcare organizations must implement continuous mechanisms for profiling, validating, cleansing, and monitoring data as it flows in and changes throughout its lifecycle. Tools offered by Melissa, for instance, can facilitate these processes across various entry points and existing data workflows, ensuring that healthcare organizations can fully trust their patient data before AI systems are engaged in clinical decision-making.

About the Author

Bob Stanley, who leads special projects at Melissa, aids clients in effectively managing the data lifecycle for insights in business, pharmaceutical, and clinical contexts. To connect with Bob, reach out via email at [email protected] or find him on LinkedIn.

The lack of patient identity control is a gap in healthcare AI.

A dimly lit hospital ward at night, an adult patient’s ID badge hovering above the bed in fragments as faint AI circuitry drips from a broken padlock connected to a sleek console beside it, bedside monitors flicker red warning lights—ominous, cinematic, dark theme.

Patient identity is the missing control in healthcare AI - Healthcare Dive

- By 2026, 81% of physicians utilize AI in healthcare, up from 38% in 2023.  
- Reliable patient data is critical for safe AI deployments to prevent potentially deadly errors.  
- Continuous data quality management is essential for accurate AI functioning and patient safety.  

In 2026, we’ve moved well beyond ‘AI is coming to healthcare.’ Adoption is accelerating, revealing the importance of reliability and governance to these new systems. While it’s important to understand what an AI model can do, it’s even more important to be 100% confident that your health system can successfully operationalize and deploy it. That, as they say, is where the rubber meets the road, and requires attention to trustworthy patient data as a foundation for AI implementations.

Ultimately, AI is only as safe as the dynamic data quality management and governance that supports it. This leads to two essential data quality questions in healthcare AI. Is the data used to build and validate the model reliable? And is the live patient data encountered by the model actually accurate, complete, current, and connected to the right patient? It’s not enough to be ‘mostly confident’ in either answer. In 12-18 months, healthcare providers and systems can expect tighter coupling between data quality programs and AI approval processes, with unsafe or low‑quality data becoming a formal reason to block or pause AI deployments.

Where we are

The AMA's March 2026 survey [1] found that 81% of physicians now use some form of AI professionally, versus 38% in 2023. To be concise, this does not mean 81% are practicing AI-assisted medicine – rather they are recognizing the value of the technology in ordinary clinical work like summarizing research and standards of care (39%), creating discharge instructions, care plans, or progress notes (30%), or documenting billing codes, charts, and visit notes (28%).

Clearly, administrative and workflow automations have already made the jump from experimentation to routine use. But there is far more to come. Generative AI is advancing in electronic health records (EHR), while predictive AI is making strides in operational areas like re-admission risk, identification of high-risk patients, scheduling and billing. Diagnostic AI, or computer vision and machine learning algorithms, has long been embedded in medical device workflows.

Ready or not

To be AI-ready in healthcare, systems need precise, accurate, dynamically cleansed, enriched and deduplicated data. Why? Deadly AI errors may emerge from poor data quality. If AI confidently orders an incorrectly dosed prescription, the outcome could be tragic. If “route of administration” metadata about a drug dose is incorrect, AI may, for example, apply a higher, potentially lethal dose of a prescribed pill.

Other examples include incorrect risk scores; misrouted, false or missed health alerts; incorrect eligibility decisions. Downstream, this can mean financial harm, erosion of trust in the system, delayed care, wrong treatment, even death.

Duplicate patient records already create substantial business and delivery-of-care problems in both healthcare research and in the clinic. (Some analysts estimate duplicate records at 8-10%. A health system maintaining a million patient records could be dealing with 80,000 duplicates – and their related effects on cost, risk, and correctness of AI-supported decisions at scale.) Fragmented records, for example if a patient record is missing potentially contraindicated supplements or medications, can create even higher risks under AI applications.

Before an AI system can reason accurately across a patient's history, the organization needs confidence that the records actually belong to that patient. Identity resolution, record matching, and deduplication can help establish a reliable, consolidated patient record – not only for safety and excellence in overall care, but before that data becomes source material for AI processes. For example, Melissa applies matching and deduplication technologies to reconcile records across systems, and create a more reliable, unified patient view.

Some of the most significant AI safeguards may occur well upstream of the AI algorithm. Melissa tools and solutions support Project US@-compliant address validation and standardization at patient registration. An initiative of ONC (Office of the National Coordinator for Health Information Technology), Project US@ is a technical specification for patient addresses, designed to prevent variations in names, addresses, and other identifying data that create fragmented or duplicate records downstream.

Data quality infrastructure is AI infrastructure

An excellent AI model can still produce a bad result if it receives an inaccurate representation of the patient. That makes something as mundane as a duplicate record, incorrect identity, outdated address, missing medication, or fragmented medical history consequential in a new way. Ultimately, the infrastructure needed for accurate patient identity and trustworthy records is the same infrastructure needed for safe AI.

Building that foundation doesn't necessarily require healthcare organizations to reinvent their data environments. A good checklist covers end-to-end data quality, starting upstream of AI and continuing through ongoing data monitoring:

Validate and standardize identity attributes at patient registration, including name, address, and contact data aligned to Project US@. Resolve and deduplicate across EHR, claims, ancillary, and affiliate feeds before those records train or trigger AI models. Profile the completeness and recency of the fields AI actually consumes, such as medications, allergies, problems, coverage, and contact information. Monitor identity breaks continuously and treat them as AI incidents, not back-office cleanup. Require data-fitness evidence in the AI approval packet alongside model-performance evidence.

AI readiness is more than treating data quality as an occasional cleanup exercise. Healthcare organizations need mechanisms to continuously profile, validate, cleanse, match, and monitor information as records enter the organization and change over time. Melissa's data quality tools, for example, support those functions across points-of-entry and existing data workflows, helping healthcare organizations establish confidence in their patient data before AI is asked to act on it.

About the Author

Bob Stanley, Melissa’s director of special projects, helps Melissa Informatics’ clients harness the entire data lifecycle for business, pharmaceutical and clinical data insight and discovery. Connect with Bob via email at [email protected] or via LinkedIn.

Patient identity is the missing control in healthcare AI

Agentic AI is making strides in the healthcare sector by improving the connection between medical decisions and patient care. Unlike basic systems that merely summarize data or answer queries, these advanced AI tools actively collect information, integrate with various software platforms, initiate actions, and pursue specific outcomes. Healthcare institutions are leveraging this technology to tackle three significant challenges: obtaining insurance approvals, ensuring patient adherence to care plans, and linking eligible individuals with clinical trials.

Navigating Prior Authorization with AI Agents

The process of prior authorization can resemble a complex relay race, with medical documentation, coverage protocols, and approval forms being shuffled among patients, providers, and payers. Any missing document can disrupt the entire procedure. An AI agent can simplify this by starting from a physician's order, determining if authorization is needed, retrieving necessary clinical evidence, and compiling the submission. It can then forward the request to the insurer, track its progress, and respond to additional information requests from the payer.

This application of AI is particularly relevant as new regulatory deadlines approach. The CMS Interoperability and Prior Authorization Final Rule mandates that affected payers implement new application programming interfaces (APIs) by January 1, 2027. These APIs will enable seamless communication, allowing providers to verify authorization requirements and manage requests and decisions electronically. An AI agent could utilize these integrations to handle the authorization process within existing provider systems and even draft appeals in case of denials, armed with patient records and payer explanations.

While clinicians will continue to assess cases requiring medical judgments, AI agents will alleviate the burden of administrative tasks such as searching for information and repetitive data entry, thereby expediting patient treatment.

Enhancing Patient Care Post-Appointment with AI Agents

After a medical appointment, patients often face a lengthy list of follow-up actions without support. Here, an agentic care navigator can seamlessly transform a discharge plan or physician referral into actionable tasks. It can arrange specialist visits, verify insurance coverage, provide preparation guidelines, and facilitate transportation. Post-discharge, the agent can offer medication reminders, track symptoms, and confirm attendance at follow-up appointments.

This innovative approach is moving beyond mere concepts into practical applications. For instance, CVS Health is unveiling Health100, an AI platform created in collaboration with Google Cloud to provide consumers with real-time, proactive support across their healthcare journey, including interactions with doctors, pharmacies, and insurance companies. Further supporting this trend, research has shown the effectiveness of automating specific phases of patient care. A clinical trial funded by the NIH demonstrated that an AI screening tool can identify hospital patients dealing with opioid use disorder and facilitate referrals to specialists.

An AI agent could further enhance this process by monitoring whether referrals result in appointments and escalating cases that are stalled or at higher risk. While it wouldn't be involved in diagnosing or altering prescriptions, its role would be to keep the patient's care plan on track and inform clinicians when human intervention is necessary.

Facilitating Access to Clinical Trials with AI Agents

Locating suitable clinical trials can be a daunting task for healthcare providers, as it requires matching a patient’s medical history, diagnoses, lab results, and genetic information against extensive eligibility criteria. Much of this critical data resides in unstructured formats within physician notes. An AI agent has the capacity to sift through these records, cross-reference findings with available trials, and generate a prioritized list of matches, each accompanied by an explanation.

A notable study published in *Nature Communications* regarding TrialMatchAI illustrates the advancements in this area. The system effectively analyzes both structured and unstructured medical records, retrieves relevant studies, and assesses eligibility criteria methodically. In real-world tests, it identified at least one applicable trial among the top 20 suggestions for 92% of cancer patients, achieving over 90% accuracy in its eligibility assessments.

A more sophisticated AI agent could go further by confirming whether a trial is currently recruiting, pinpointing the nearest site, and preparing patient information for review. It could also alert research coordinators when new test results affect a patient's eligibility. Ultimately, researchers and clinicians would retain control over the final decision-making regarding screening, consent, and enrollment. The primary advantage of an AI agent lies in uncovering opportunities that busy medical teams or patients might overlook.

In the short term, the true promise of agentic AI in healthcare shines where coordination often falters. The challenges of prior authorization, patient navigation, and trial matching all involve multiple systems, extensive follow-ups, and the careful handling of sensitive information. With stringent controls and human oversight, these AI agents have the potential to bridge administrative gaps, helping patients access approved care, follow-up services, and experimental treatments more efficiently.

AI Healthcare Agents Speed Up Patient Access to Care

A bright, sunlit modern clinic reception: a smiling adult patient leans on the counter while a medical professional holds a sleek tablet displaying a friendly AI-powered care plan interface—colorful icons, progress bars, and instant appointment confirmations. Warm natural light streams through large windows onto minimalist potted plants and bustling staff guiding other adults to consultation rooms in the background.

Healthcare AI Agents Help Patients Get Care Faster - PYMNTS.com

- Agentic AI enhances healthcare by streamlining insurance approvals, patient tracking, and clinical trial matching.
- AI agents automate processes like prior authorization, reducing staff workload and delays in treatment.
- These systems effectively match patients to clinical trials, improving access to relevant treatment opportunities.

Agentic AI is entering healthcare through the work that happens between a medical decision and the care a patient eventually receives. These systems go beyond summarizing records or answering requests. They can collect information, move between software systems, initiate an action and continue working toward a defined result. Healthcare organizations are applying that capability to three persistent problems: securing insurance approval, keeping patients on track and connecting eligible people with clinical trials.

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions .

Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

AI Agents Take On Prior Authorization’s Paper Chase

Prior authorization operates like a relay race in which medical records, coverage rules and approval forms pass among providers, payers and patients. A missing document can stop the entire process. An AI agent can start with a physician’s order, determine whether approval is required, retrieve the relevant clinical evidence and assemble the submission. It can then send the request to the insurer, monitor its status and respond when the payer asks for more information.

The regulatory calendar is making this use case especially timely. Under the CMS Interoperability and Prior Authorization Final Rule, affected payers generally must meet new application programming interface requirements by Jan. 1, 2027. Those APIs will let providers determine authorization requirements and exchange requests and decisions electronically. An agent could use those connections to manage the process from inside a provider’s existing system. It could also draft an appeal after a denial, supported by the patient’s records and the payer’s stated reason.

Clinicians would still review cases requiring medical judgment. The practical gain comes from removing the searches, status checks and repetitive data entry that consume staff time and can delay treatment.

AI Agents Keep Patients Moving After the Appointment

Healthcare often assigns patients a long list of next steps, then leaves them to coordinate the work. An agentic care navigator could turn a discharge plan or physician referral into a series of completed tasks. It could schedule a specialist visit, check insurance coverage, send preparation instructions and arrange transportation. After discharge, the agent could deliver medication reminders, ask about symptoms and confirm that the patient attended follow-up appointments.

We’d love to be your preferred source for news. Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks! Add as Preferred Source

This model is beginning to move from concept to product. CVS Health is launching Health100, an AI-based platform developed with Google Cloud that is designed to give consumers proactive, real-time support across doctors, pharmacies and insurance providers. Research has also demonstrated the value of automating narrower steps in the care journey. An NIH-supported clinical trial found that an AI screening tool could help identify hospital patients with opioid use disorder and generate referrals to specialists.

An agent could extend that approach by tracking whether the referral led to an appointment and escalating stalled or high-risk cases. The agent wouldn’t diagnose a patient or change a prescription. Its job would be to keep the care plan moving and alert a clinician when human intervention is needed.

AI Agents Find Clinical Trials Patients Might Otherwise Miss

Finding an appropriate clinical trial can require clinicians and research coordinators to compare a patient’s diagnosis, treatment history, laboratory results and genetic markers against lengthy eligibility criteria. Much of that information sits in physician notes or other unstructured records. An AI agent can search those records, compare the findings with available trials and produce a ranked list of potential matches with an explanation for each recommendation.

A 2026 Nature Communications study of TrialMatchAI shows how far this capability has progressed. The system processes structured records and unstructured physician notes, retrieves relevant studies and conducts a criterion-by-criterion eligibility assessment. In real-world validation, it found at least one relevant trial among the top 20 recommendations for 92% of oncology patients. Experts validated more than 90% accuracy in its eligibility classifications.

An agentic version could take additional steps, including checking whether a trial is recruiting, identifying the closest participating site and preparing the patient’s information for review. It could also notify the research coordinator when new test results change a patient’s eligibility. Researchers and clinicians would retain authority over final screening, consent and enrollment. The agent’s value would come from finding opportunities that time-pressed medical teams or patients may never discover.

Agentic AI’s near-term healthcare value is strongest where coordination breaks down. Prior authorization, patient navigation and trial matching each require multiple systems, repeated follow-up and careful handling of sensitive information. With firm controls and human review, agents could reduce those administrative gaps while helping patients reach approved care, follow-up services and experimental treatments faster.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

Healthcare AI Agents Help Patients Get Care Faster

In the evolving landscape of artificial intelligence, a healthcare system in the Bay Area is harnessing this technology to innovate care delivery and support medical professionals. Sutter Health, a nonprofit organization headquartered in Sacramento, caters to over 3.5 million patients across Northern and Central California and has notably enhanced its AI integration in clinical settings over recent years.

In 2024, the organization inaugurated its San Francisco “innovation center,” aiming to collaborate with AI and tech firms on developing digital health applications, remote monitoring solutions, and other advanced technologies.

Sutter Health plays a pivotal role as the West Coast hub for clinical AI advancements by partnering with Aidoc, a medical technology company. This collaboration began last year, focusing on utilizing Aidoc’s AI system to expedite disease detection and improve patient care. In January, their partnership expanded to the field of radiology, wherein the AI technology was employed to analyze imaging scans and generate initial drafts of radiology reports.

Additionally, in February, Sutter Health teamed up with OpenEvidence, an AI-driven medical search engine designed to assist Sutter’s healthcare professionals in quickly finding evidence-based answers to clinical inquiries.

This spring, Sutter Health publicly announced its intention to collaborate with Santa Clara University to establish the Bay Area’s first new medical school in over a century. This institution is set to emphasize emerging AI technologies, aiming to enhance the efficiency and accuracy of the medical profession.

Ashley Beecy, the organization’s inaugural Chief AI Officer, recently discussed with Bay Area News Group how Sutter Health is implementing artificial intelligence to aid doctors, enhance patient experiences, and revolutionize healthcare services.

In her capacity as Chief AI Officer, Beecy leads initiatives focused on responsibly integrating AI into patient care while optimizing service delivery for both patients and providers. Her team's focus is on developing tools and forming partnerships to enhance healthcare operations, providing training for employees on AI utilization, and instituting governance to ensure the safe application of AI technologies.

Beecy emphasized that AI is essential for easing the workload of clinicians and ensuring the highest quality of patient care. Situated in Northern California, a hub for AI development, Sutter Health recognizes the transformative potential of these technologies.

The integration of AI in clinical settings allows Sutter Health’s healthcare providers to utilize the vast amounts of data present in electronic health records more effectively. For instance, AI tools assist in synthesizing this data to provide valuable insights during patient interactions, enabling clinicians to prioritize patient engagement over administrative responsibilities. Innovations such as chart summarization tools and ambient scribing systems significantly reduce the documentation burden, allowing healthcare professionals to focus on patient care and return to work refreshed.

On the patient side, Sutter Health has introduced a user-friendly chatbot that grants patients access to their electronic health records at any time. This initiative enables patients to inquire about new symptoms or determine the appropriate care provider while ensuring data privacy and personalized access.

To address concerns about AI, Sutter Health is committed to building trust through transparency and education. Beecy stated that demonstrating the evaluation, monitoring, and reporting of AI tools before they are rolled out helps to alleviate skepticism among patients and staff alike.

Overall, Beecy noted that AI innovations could enhance the healthcare sector by reducing administrative tasks for care teams, allowing them to concentrate on the human aspects of patient interactions and complex clinical decision-making. The goal is to make healthcare access as seamless as possible, mirroring the convenience often seen in other industries like travel and retail.

With a background in computer systems engineering and a medical degree complemented by additional training, Ashley Beecy is well-equipped to spearhead these initiatives. Based in Emeryville, CA, her multifaceted expertise positions her to navigate the intersection of technology and healthcare effectively.

Bay Area healthcare system balances tech upgrades and patient trust while navigating AI.

A sunlit Bay Area clinic reception bathed in warm morning light, where a doctor and adult patient sit at a sleek white desk. The doctor holds a tablet showing a softly glowing AI dashboard with pastel network lines, while a tech specialist nearby adjusts controls on a modern touchscreen stand—capturing the harmony of human care and advanced technology.

How a Bay Area healthcare system is navigating AI by balancing tech upgrades and patient trust - The Mercury News

- Sutter Health is leveraging AI to enhance clinical care and streamline workflows for doctors.  
- The healthcare system launched an innovation center to develop digital health technologies and applications.  
- Patients can access a chatbot for personalized health records and connect with care providers.  

As artificial intelligence begins to transform the workforce and reshape industries, one Bay Area healthcare system is using the technology to support doctors and streamline care.

Sutter Health, a not-for-profit healthcare system based in Sacramento, serves more than 3.5 million patients throughout Northern and Central California and has significantly expanded its use of artificial intelligence in clinical care in recent years.

In 2024, the healthcare giant launched its San Francisco “innovation center” to work with AI and tech firms on digital health apps, remote monitoring programs and other emerging technologies.

Sutter Health serves as the West Coast hub for clinical AI innovation for the medical technology company Aidoc. The healthcare system partnered with Aidoc last year to use its AI operating system to more quickly detect disease and provide patients with care.

In January, Sutter Health expanded the partnership to include radiology, using Aidoc’s AI technology to review and flag patients’ imaging scans and produce preliminary drafts of radiology reports.

Sutter Health also partnered with OpenEvidence, an artificial intelligence platform and medical search engine, in February to help Sutter physicians find fast, evidence-based answers to clinical questions.

And, this spring, Sutter Health announced plans to launch the Bay Area’s first new medical school in over a century, in partnership with Santa Clara University, with plans to center the school around emerging AI technologies and make the medical profession more efficient and accurate.

Sutter Health’s inaugural Chief AI Officer, Ashley Beecy, spoke with Bay Area News Group about some of the ways Sutter Health is using artificial intelligence to support physicians, improve patient care and transform the healthcare industry.

Q: What is your role as Chief AI Officer for Sutter Health?

A: As Chief Artificial Intelligence Officer at Sutter Health, I lead efforts to use AI responsibly to improve care and make health care work better for patients and clinicians. My team really is the Center for Applied AI.

Our work can include building tools or partnering with vendors where there’s opportunity. We support training employees about AI and how to use it in their own work. We also lead AI governance. That includes making sure that we’re using AI safely and we can monitor it appropriately.

Q: Why is AI a priority for Sutter Health?

A: At Sutter, we want to make sure that we can make work easier for our clinicians and we also deliver the best care possible for our patients. And AI is really an enabler to be able to do that. We’re in Northern California — one of the largest AI centers worldwide.

Q: How are Sutter Health clinicians using AI?

A: When we think about making work easier for our care teams, we think about all the digitized data in the electronic health record. How do we synthesize that in a way that they have better insights when they’re in an encounter with a patient? Or when they’re in that patient encounter, how can they focus on that human touch with the patient and not the administrative burden of taking notes and documenting it? So we have chart summarization tools and tools that can summarize information to help doctors prepare for a visit. We also have ambient scribes to be able to document the encounter so that they don’t have to do that when they go home; they can spend more time with their families, and they come to work the next day really refreshed to engage in patient care again.

Q: How are Sutter Health patients using AI?

A: We just rolled out a patient-facing chatbot where our patients can connect any time of day to their electronic health record and ask questions against it. Whether they’re having new symptoms or they want to understand who’s the right provider or care team member to see, they can go to a place that has the right data privacy protections and connects them to their own personalized data as opposed to using the publicly available tools.

Q: How is Sutter Health navigating distrust and skepticism around AI?

A: With every digital tool, it’s about building trust with our patients, our care team members and our employees. We do that by showing them that we’re evaluating these tools before we deploy, that we’re being transparent about the performance and how to use them, that we’re onboarding people and teaching them how to use it, that we’re monitoring over time and we’re reporting back the outcomes and safety measurements that we establish in advance of any deployment.

Q: What do innovations in AI mean for the healthcare industry overall?

A: It means that we’re actually increasing the human touch and care by taking some of the administrative work off of our care teams. We’re allowing them to focus on that human engagement with the patient, to focus on the really complicated cases that require their human-level thinking and intelligence applied to clinical decision-making.

One of our goals is to make sure that we provide the right access to our patients. So anybody that wants to come to our health system, we make it as easy as possible for them to find the right care team. We know that as a consumer in other industries, there’s a lot of technology to support things you do, whether it’s travel or shopping. And in our mind, there’s no reason that healthcare shouldn’t parallel these industries in creating the same level of experience.

Ashely Beecy Profile

Position: Chief Artificial Intelligence Officer at Sutter Health

Education: Bachelor’s degree in computer systems engineering at Boston University, medical degree from Albert Einstein College of Medicine in New York City and additional training at Columbia University and Technische Universität Dresden. She is board certified in internal medicine, cardiovascular medicine, echocardiography and clinical informatics.

Residence: Emeryville, CA

5 things to know about Ashley Beecy

How a Bay Area healthcare system is navigating AI by balancing tech upgrades and patient trust

Corporate earnings in China are experiencing significant growth, providing valuable insights for investors seeking lucrative stock opportunities, as highlighted in a recent Goldman Sachs report. Their analysis, published on September 7, indicates that nearly half of the constituents in the MSCI China index surpassed earnings expectations in the second quarter, with technology and healthcare sectors leading the charge. The report reveals that earnings for Chinese stocks tracked by Goldman Sachs increased at a remarkable 24% year-over-year during this quarter, marking the highest growth rate in five years and a notable rise from just 6% in the prior quarter.

A substantial portion of this growth can be attributed to companies involved in artificial intelligence, mirroring trends observed across global stock markets. However, as AI stocks become increasingly saturated, many investors are now seeking emerging growth prospects outside of the traditional AI-driven tech landscape. 

The report notes that discussions among management and investors from over 1,500 Chinese firms have expanded beyond just hardware and semiconductors to include downstream/application sectors like data centers, AI models, automotive, and healthcare. To identify promising investments, Goldman’s analysts filtered their list of buy-rated companies for those projected to achieve over 15% annual earnings growth through 2027, as well as those that have recently seen a median increase of 7% in earnings per share estimates. 

Among the final selections, the pharmaceutical sector prominently featured, representing one-third of the 12 companies identified. For example, Suzhou-based Innovent Biologics is anticipated to see its earnings more than double over the next year, with Goldman’s estimates significantly surpassing market consensus. This company boasts a robust pipeline focused on cancer and metabolic disease treatments. Similarly, Shanghai-listed BeOne Medicines, which also operates on Nasdaq and in Hong Kong, is projected to double its earnings within the same timeframe, specializing in cancer therapies. 

CSPC, another Hong Kong-listed entity, expects 26% earnings growth, primarily focused on drugs for the nervous system and cardiovascular health. Hansoh Pharma, also listed in Hong Kong, has a forecasted earnings increase of 15%, with a portfolio that includes metabolic and cancer treatments. Should these companies meet their earnings projections, they would outperform Goldman’s broader expectation of just 8% earnings growth for the MSCI China index, which is significantly more conservative than the consensus prediction of 17%. 

The insights provided by Goldman Sachs suggest a promising landscape for Chinese stocks, particularly within the healthcare sector, as investors look for potential high-return opportunities amid a shifting market dynamic.

Goldman selects Chinese healthcare stocks for a post-AI investment strategy.

A sleek, sunlit trading room: morning light streams through a window onto a curved display bearing a prominent Goldman Sachs logo. Beside it, a gently unfurled Chinese flag. Multiple monitors show rising biotech and healthcare stock charts overlaid with a subtle DNA helix and a stethoscope icon. Two sharply dressed traders lean in over a polished desk, deep in discussion about the post-AI healthcare play.

Goldman picks China healthcare stocks for a post-AI trade - CNBC

- Nearly half of MSCI China index stocks beat earnings estimates in Q2 2023.  
- Chinese corporate earnings surged 24% in Q2, driven largely by AI-related companies.  
- Goldman Sachs identified healthcare stocks as major opportunities for future growth potential.  

Chinese corporate earnings are on a tear, giving investors clues on which stock sectors offer untapped returns, according to Goldman Sachs. Nearly half of MSCI China index constituents beat estimates in the second quarter, with IT and healthcare stocks dominating, the firm's portfolio strategy research team said in a Sept. 7 report. Chinese stocks' earnings tracked by Goldman Sachs grew at their highest quarterly pace in five years at 24% in the second quarter from a year ago — accelerating from just 6% in the first quarter. Most of the gains were driven by artificial intelligence-related names, as has been the case in other stock markets globally. "As AI-related stocks become increasingly crowded and concentrated, many investors are looking for additional growth opportunities beyond the AI Hard Tech ecosystem," they said. Earnings call transcripts of more than 1,500 Chinese companies revealed management and investor "discussions have expanded beyond hardware and semiconductor sectors to downstream/application sectors, such as data center operators, AI models, autos, and healthcare," the report said. To find potential winners, the analysts screened their buy-rated coverage for Chinese companies with expected earnings growth of more than 15% annually through 2027, and an increase in earnings per share estimates by a median of 7% over the past month. The screen favored stocks where Goldman's earnings estimates were above consensus, indicating how earnings could be revised higher soon. Healthcare, particularly pharmaceutical stocks, showed up prominently, accounting for one-third of the 12 names that made the final list: Suzhou-based Innovent Biologics' earnings are expected to more than double in the year ahead, while Goldman's consensus is 54 percentage points above the consensus. The Hong Kong-listed stock claims a pipeline of drugs including cancer and metabolic diseases. Shanghai-listed BeOne Medicines earnings are also expected to more than double in the year ahead. The company, which develops cancer treatments, is also listed on the Nasdaq and Hong Kong. Hong Kong-listed CSPC earnings are expected to grow by 26%. The mainland Chinese company's major businesses include nervous system and cardiovascular drugs. Hong Kong-listed Hansoh Pharma earnings are forecast to grow by 15%. The company's products cover metabolic treatments as well as cancer. If earnings for each of the four companies grow as forecast, that would be above the investment firm's expectation for the broader Chinese stock market. The Goldman analysts expect MSCI China earnings to grow by 8% this year, far more conservative than the consensus forecast for 17%. — CNBC's Michael Bloom contributed to this report.

Goldman picks China healthcare stocks for a post-AI trade

For years, artificial intelligence (AI) has been anticipated to revolutionize the field of medicine, yet a recent thorough review indicates that we may be at a pivotal moment where technological progress significantly outpaces its clinical application. Researchers Deepika Yadav, Pooja Yadav, and Hemant Yadav conducted an extensive analysis, published in Discover Informatics, aiming to chart how various AI technologies, including machine learning, deep learning, the Internet of Things (IoT), and blockchain, are utilized within healthcare and to identify prevalent gaps.

Their findings reveal a sobering reality: many healthcare AI systems currently outlined in literature seem to be stuck at intermediate stages of development, often only validated in labs or pilot projects rather than being routinely implemented in clinical settings. 

Employing the PRISMA framework—deemed the gold standard for systematic reviews—the researchers began with a pool of 1,860 scholarly articles sourced from leading databases like PubMed and IEEE Xplore. After carefully eliminating duplicates and scrutinizing titles and abstracts, they narrowed it down to 331 articles for a detailed review, ultimately including 108 studies that focused on literature published between 2021 and 2026 to highlight the latest in AI healthcare advancements. Each study was assessed using the Technology Readiness Level (TRL) framework, originally created by NASA, which categorizes technologies based on their maturity and readiness for deployment.

The TRL analysis shed light on one of the review's key conclusions: a significant majority of healthcare AI tools are in the TRL 3 to TRL 5 range. This indicates that they remain at the conceptual phase or as early prototypes and have not transitioned to large-scale clinical implementation, which aligns with TRL 7 through 9. This suggests that AI in healthcare is still largely in a state of flux, transitioning from theoretical applications to tangible clinical use, with TRLs 4 and 5 being the most prevalent maturity stages found.

The review categorizes AI applications in the healthcare sector into an organized hierarchy. It details how machine learning, which develops algorithms for data analysis, is already employed in tasks such as monitoring patient facial expressions and diagnosing illnesses via techniques like Support Vector Machines. On the other hand, deep learning—characterized by its use of advanced neural networks—has made significant strides, particularly in medical imaging. For example, systems like Google's DeepMind and IBM's Watson have achieved results in identifying malignant tumors that rival those of experienced radiologists.

The scope of applications described in the review is extensive. Deep learning has transformed medical image analysis, which is crucial across various fields such as radiology and dermatology, while predictive models assist in risk assessments to improve patient safety and reduce errors. Additionally, AI facilitates faster diagnostics via automated screening and enhances treatment plans while addressing healthcare fraud through blockchain technology, thereby improving patient engagement and preemptive care models.

Moreover, the review discusses the fusion of AI with other advanced technologies. The IoT enables real-time data sharing from wearable devices, paving the way for remote monitoring and telehealth innovations, especially when combined with 5G networks. Blockchain, primarily known for its role in cryptocurrencies, secures electronic medical records, enhancing integrity and interoperability across systems. There are studies outlined that explore the integration of AI and blockchain for managing health records along with technologies that create digital twins tailored for individual patients.

Nevertheless, the report underscores significant challenges facing the field. Data collection issues impede progress, as patient confidentiality concerns and stringent regulations such as GDPR complicate data access and research collaboration. Consequently, the quality of data remains inconsistent, adversely affecting AI performance. Algorithm-related challenges, including bias in training data and the black-box phenomenon—which renders AI outputs incomprehensible even to developers—raise serious questions about clinical accountability and the reliability of AI-assisted recommendations.

Ethical dilemmas further complicate the technological landscape. Ascribing liability for AI-driven errors is challenging due to the opaque nature of AI decision-making. Additionally, while regulatory bodies like the FDA work toward establishing frameworks for evaluating healthcare AI, concerns regarding job security among healthcare professionals persist, emphasizing the need for role evolution rather than job loss in the face of AI advancements. The transition from AI research to real-world application is fraught with barriers, such as limited representation in study populations and the necessity for integration that seamlessly fits into existing healthcare workflows. There is also a notable gap in research focused on mental health, chronic conditions, and elder care, and the synergy between AI, IoT, and blockchain remains largely unexplored.

In light of these findings, the authors lay out a clear agenda for the future of medical AI. They highlight the urgent need for clinical validation and effective deployment of AI tools, advocating for research that focuses on explainable AI to foster clinician trust, federated learning for secure data sharing, and the use of large language models for decision support and data synthesis. They also propose advancements in digital twin technologies to enhance individualized patient care while recommending stricter regulatory measures and extensive clinical trials to ensure safe integration into healthcare practices. The overarching message conveys a cautious optimism: while AI systems have already shown considerable promise in improving diagnostics and healthcare delivery, significant efforts are needed to bridge the gap between research efficacy and practical clinical application.

AI in Healthcare Set to Revolutionize Medicine, Yet Many Tools Remain Ineffective

Scene in a bright hospital corridor bathed in warm morning light: a sleek robotic arm holding a stethoscope stands beside a wall-mounted screen showing a half-lit neural network diagram. The neural lines on one side glow vivid blue, the other side flickers off. A doctor in a white coat studies notes on a clipboard, brow furrowed, while thick cables coil around the robot’s base, anchoring it to the floor—a visual of promise and stalled progress.

AI in Healthcare Poised to Transform Medicine, But Most Tools Still Stuck in the Lab - Bioengineer.org

- A systematic review indicates healthcare AI technologies are still not clinically implemented.  
- Most AI systems are at intermediate maturity levels, lacking large-scale deployment evidence.  
- Data privacy and algorithmic bias challenge effective integration of AI in healthcare practices.  

Artificial intelligence has promised to remake medicine for decades, but a sweeping new systematic review suggests the field is at a decisive turning point: the technology is advancing faster than its ability to reach the clinic. Researchers Deepika Yadav, Pooja Yadav and Hemant Yadav, publishing in the journal Discover Informatics, have compiled one of the most comprehensive maps to date of how machine learning, deep learning, the Internet of Things and blockchain are being deployed across healthcare, and where the gaps remain. Their conclusion is striking. Most healthcare AI systems described in the recent literature are still trapped at intermediate stages of technological maturity, validated in laboratories or pilot settings rather than in routine clinical practice.

The team applied the PRISMA framework, the gold-standard protocol for systematic reviews, to sift through the scientific record. Beginning with 1,860 records drawn from major databases including PubMed, IEEE Xplore, Scopus, ScienceDirect, ACM, Springer and Wiley, they removed 657 duplicates and screened 1,203 titles and abstracts. After full-text eligibility assessment of 331 articles, 108 studies were ultimately included in the qualitative and quantitative synthesis. The review focused on literature published between 2021 and 2026, capturing the most recent wave of AI-driven healthcare innovation. Each selected study was then evaluated using the Technology Readiness Level framework, a nine-stage maturity scale originally developed by NASA, to determine how close each technology actually is to real-world deployment.

The TRL analysis produced one of the review’s most consequential findings. The majority of healthcare AI technologies cluster between TRL 3 and TRL 5, meaning they exist as conceptual frameworks, proof-of-concept demonstrations, laboratory prototypes or early tests in relevant environments. Very few studies demonstrated large-scale clinical implementation or fully operational deployment, which would correspond to TRL 7 through 9. In practical terms, the review shows that healthcare AI remains largely in transition from research-based proof-of-concept systems to genuine clinical applications, with TRL 4 and TRL 5 being the most common maturity levels observed across the literature.

Technically, the review organizes the field into a hierarchical taxonomy of AI methods and their healthcare applications. Machine learning, the older sibling of the AI family, develops data-analysis algorithms that extract features from data and improve with exposure to more examples. Techniques such as Support Vector Machines and Naïve Bayes classifiers are already being used to classify facial expressions for patient monitoring and disease diagnosis. Deep learning goes further, employing artificial neural networks with multiple hidden layers and millions or even billions of parameters. These architectures have proven so powerful in medical imaging that systems such as Google’s DeepMind and IBM’s Watson have demonstrated performance on malignant tumor detection that rivals or exceeds human radiologists, according to studies cited in the review.

The application landscape the authors map is remarkably broad. Medical image analysis, transformed by deep learning, underpins modern diagnosis, treatment and monitoring across radiology, pathology, dermatology and ophthalmology. Disease prediction and risk assessment models help clinicians anticipate dangers, identify lesion locations and reduce medical errors. Automated screening systems are accelerating diagnostics through predictive analytics, medical imaging and clinical decision support. Beyond diagnosis, AI is reshaping treatment planning and patient care, detecting healthcare insurance fraud through blockchain-empowered analytics, powering patient engagement tools that generate personalized insights, and enabling preventive care through predictive models that flag disease risks before symptoms appear. Subdomains such as robotic surgery, virtual health aides, drug discovery and remote patient monitoring round out the picture of a technology touching nearly every corner of medicine.

The review also highlights how AI is converging with other emerging technologies. The Internet of Things connects wearable devices and sensors that stream continuous patient data, enabling real-time telehealth, remote diagnosis and even remote surgery when paired with 5G networks. Blockchain, the decentralized and immutable ledger technology originally conceived for Bitcoin, offers integrity, traceability and non-repudiation for electronic medical records, securing data sharing across institutions through hash chains, digital signatures and consensus mechanisms. The authors point to studies combining AI and blockchain for secure health record management and patient identity systems, as well as digital twin technologies that create patient-specific computational models for personalized medicine, including neurosymbolic digital twins for cardiovascular disease prediction.

Yet the challenges catalogued in the review are formidable. Data collection remains a fundamental bottleneck: patient confidentiality concerns limit the availability of relevant information, and privacy regulations such as GDPR, while essential for protecting personal data, complicate research collaboration. Data quality problems, including inconsistent records, directly degrade algorithm performance. On the algorithmic side, bias in training data can distort AI outcomes, and overfitting causes models to latch onto irrelevant correlations. The notorious black-box problem, in which deep learning systems reach conclusions that even their creators cannot fully explain, undermines clinical trust and accountability. When a physician cannot understand why an AI system recommends a treatment, the reliability of medical advice itself comes into question.

Ethical and social concerns compound the technical ones. Accountability for AI errors is difficult to assign when decision-making is opaque, and universal ethical standards for healthcare AI have yet to be established, although regulatory bodies such as the FDA are developing assessment frameworks. Fear of job displacement fuels skepticism among healthcare workers, and the authors argue that roles must be transformed rather than eliminated to allow AI advancement. Clinical implementation poses its own barriers: most AI research has not been developed within actual clinical settings, generalization to diverse patient populations is complicated by small or biased training datasets, and successful adoption requires stakeholder engagement, workflow integration that does not disrupt care, and training for healthcare personnel. The review also flags that mental health, chronic disease and elder care remain understudied areas, and that integration of AI with IoT and blockchain is still restricted, leaving fertile ground for future research.

The authors chart a forward agenda that reads like a roadmap for the next decade of medical AI. Clinical validation and real-world deployment emerge as the most critical research needs, given that most current studies sit at intermediate readiness levels. Future work should prioritize explainable AI models that enhance transparency and clinician trust, federated learning that enables privacy-preserving collaboration across institutions, and large language models and generative AI for clinical decision support, medical knowledge extraction and synthetic data generation. Digital twin technologies could personalize care at the level of the individual patient, while stronger regulatory frameworks and extensive clinical validation trials are needed to guarantee safe, effective and sustainable integration. The review’s ultimate message is one of measured optimism: AI-based systems have already delivered marked improvements in diagnostic precision, tailored treatment strategies and healthcare efficiency, but building trustworthy, scalable and clinically applicable systems will require the field to close the gap between what works in the laboratory and what works at the bedside.

Subject of Research: A PRISMA-based systematic review of artificial intelligence applications, techniques and challenges in healthcare

Article Title: A PRISMA-based Systematic Review of Artificial Intelligence in Healthcare, its Applications and Challenges

Article References: Yadav, D., Yadav, P., & Yadav, H. (2026). A PRISMA-based Systematic Review of Artificial Intelligence in Healthcare, its Applications and Challenges. Discover Informatics, 1(1), Article 12. https://doi.org/10.1007/s44564-026-00009-y

Image Credits: AI Generated

DOI: 10.1007/s44564-026-00009-y

Keywords: Artificial Intelligence, Healthcare, Machine Learning, Deep Learning, IoT, Blockchain, Technology Readiness Level, PRISMA, Medical Imaging, Explainable AI, Data Privacy, Algorithmic Bias

Cite Scienmag News

APA MLA Chicago

Blake Davidson. (September 12, 2026). AI in Healthcare Poised to Transform Medicine, But Most Tools Still Stuck in the Lab. Scienmag. https://scienmag.com/ai-in-healthcare-poised-to-transform-medicine-but-most-tools-still-stuck-in-the-lab/

Blake Davidson. “AI in Healthcare Poised to Transform Medicine, But Most Tools Still Stuck in the Lab.” Scienmag, 12 September 2026, https://scienmag.com/ai-in-healthcare-poised-to-transform-medicine-but-most-tools-still-stuck-in-the-lab/. Accessed 12 September 2026.

Blake Davidson. “AI in Healthcare Poised to Transform Medicine, But Most Tools Still Stuck in the Lab.” Scienmag. September 12, 2026. https://scienmag.com/ai-in-healthcare-poised-to-transform-medicine-but-most-tools-still-stuck-in-the-lab/

Copy citation Download RIS

Tags: AI deployment in clinical practiceAI in healthcarealgorithmic biasArtificial Intelligenceblockchainblockchain healthcare solutionschallenges in translating healthcare AIData Privacydeep learningdeep learning for clinical applicationsexplainable AIgaps in AI healthcare implementationHealthcarehealthcare AI technological maturityInternet of Things in healthcareIoTMachine learningmachine learning in medicinemedical AI validation stagesMedical ImagingPRISMArecent advancements in healthcare AIsystematic review of AI healthcare toolsTechnology Readiness Level

AI in Healthcare Poised to Transform Medicine, But Most Tools Still Stuck

Healthcare professionals require a clearer understanding of their responsibilities regarding artificial intelligence (AI) usage, according to a recent governmental report.

The document, titled ‘National Commission into the Regulation of AI in Healthcare: Recommendations for a Future Regulatory Framework’, was released on September 10, 2026. It advocates for regulatory measures that mandate standards or oversight for devices that present a significant risk to patients.

The report highlights current negligence claim patterns, suggesting that healthcare professionals and providers are more frequently targeted due to their defined duty of care. It warns that accountability for mistakes might fall on these professionals without adequately acknowledging the role of AI systems and the broader design of healthcare systems.

Mark DasGupta, who oversees professional development at the Pharmacists’ Defence Association (PDA), pointed out that many of the issues raised in the National Commission’s report align with previously identified concerns in their guidance on the safe and effective implementation of AI in pharmacy practice, released in August 2026.

He noted that as AI becomes more embedded within healthcare and pharmacy operations, there is an increasing need to address accountability, responsibility, and liability. While pharmacists acknowledge that AI can affect decision-making, there remains uncertainty about how responsibility is distributed among healthcare professionals, employers, organizations that implement AI, and developers.

“One significant concern for the PDA is that pharmacists might still be held professionally accountable for decisions influenced by AI systems, even when they have limited insight or control over those systems,” DasGupta explained.

He added that a unified approach to enhance professional capability would be advantageous in establishing consistent standards and understanding across various healthcare professions. This initiative should support rather than replace existing profession-focused guidelines and training.

The report's Recommendation 32 calls for a collaborative approach involving stakeholders such as the Department of Health and Social Care, regulatory bodies, and royal colleges to advance AI integration in healthcare. This inclusive strategy should encompass initial education, postgraduate training, and ongoing professional development. It also advises that contracts between AI manufacturers and healthcare providers should explicitly define the allocation of responsibilities for risk management.

Tase Oputu, president of the Royal College of Pharmacy, emphasized the necessity for clear accountability among developers, suppliers, commissioners, healthcare providers, and professionals. He stated that while pharmacy professionals should be responsible for decisions within their control, risk responsibilities—like those stemming from product defects or organizational failures—should be allocated according to who can best manage those risks throughout the AI lifecycle, rather than automatically burdening frontline clinicians.

Oputu expressed agreement on the need for practical guidance to assist healthcare professionals in comprehending their responsibilities regarding AI use. He highlighted ongoing discussions with stakeholders about these challenges during digital and AI roundtables.

Reflecting on Recommendation 32, Oputu reiterated that pharmacy professionals require appropriate education, training, and opportunities to acquire the skills necessary for safe and effective AI utilization. He added that their existing policy recognizes the value of enhancing digital and AI skills throughout the pharmacy workforce, with educational programs tailored to varying roles and levels of AI involvement.

Kathie Cashell, chief executive of the General Pharmaceutical Council (GPhC), welcomed the findings of the report, stating that the GPhC would carefully evaluate its recommendations. She agreed on the necessity of a coordinated approach to foster professional capability for the responsible use of AI in healthcare. Moving forward, the GPhC plans to collaborate closely with other regulatory agencies and health organizations to promote safe and effective AI integration.

Recently, the GPhC has released resources concerning AI application in pharmacy practice, alongside guidance on revalidation and training. They intend to undertake additional initiatives as technology evolves, including updates to the standards applicable to pharmacists and pharmacy technicians.

National Commission calls for clear guidelines on AI responsibilities in healthcare.

A sunlit hospital conference room where a diverse panel of doctors and commission members sit around a polished wooden table strewn with folders labeled “AI in Healthcare.” On the far wall, a large screen gently glows with an outline of a human silhouette and overlaid bullet points titled “AI Responsibilities” in crisp green text. A small desk placard reads “National Commission.” Stethoscopes drape over chairs, soft daylight filters through tall windows, and the mood is hopeful and focused as they seek clarity.

Clarity on AI responsibilities in healthcare needed, National Commission says - The Pharmaceutical Journal

- Healthcare professionals need clearer guidelines on responsibilities regarding AI integration in healthcare.  
- Current negligence procedures unfairly hold healthcare providers accountable for AI-related errors.  
- A coordinated approach is essential for developing AI standards and training in healthcare.

Healthcare professionals need clarity on their responsibilities when using artificial intelligence (AI), a government report has said.

The report — titled ‘National Commission into the regulation of AI in healthcare: recommendations for a future regulatory framework‘ — published on 10 September 2026, also recommended that regulations require standards or oversight for devices “where there is sufficient risk to patients”.

Under current negligence procedures, claims are disproportionately likely to be brought against healthcare professionals and healthcare providers because these have the clearest duty of care, the report suggested.

“Responsibility for errors may be transferred onto healthcare professionals and providers without sufficiently recognising the influences of AI systems and wider system design,” it warned.

Mark DasGupta, director of professional development at the Pharmacists’ Defence Association (PDA), said that many issues highlighted by the National Commission report “are consistent with concerns that the PDA has already identified” in guidance on the safe and effective use of AI in pharmacy practice, published in August 2026.

“As AI becomes increasingly integrated into healthcare systems and pharmacy workflows, questions about accountability, responsibility and liability are being raised more frequently. Pharmacists recognise that AI may influence decisions, but there remains uncertainty about how responsibility is shared between healthcare professionals, employers, organisations deploying AI systems and the developers of those systems,” he said.

“One of the key concerns identified by the PDA is that pharmacists may remain professionally accountable for decisions influenced by systems over which they have limited visibility or control.”

DasGupta added: “A coordinated approach to building professional capability could be beneficial in helping to establish consistent standards and understanding across healthcare professions. However, it should complement, rather than replace, profession-specific guidance and training.”

Recommendation 32 of the National Commission’s report said that stakeholders — including the Department of Health and Social Care and devolved health departments, regulators and royal colleges — should take a coordinated approach to developing the use of AI in healthcare.

It added that the approach should include initial education, postgraduate training and continuing professional development. Contracts between AI device manufacturers and healthcare providers should also include an explicit allocation of responsibilities for controlling risk, the report recommended.

Tase Oputu, president of the Royal College of Pharmacy, said: “In our contribution to the National Commission’s work, we highlighted the need for clear accountability across developers, suppliers, commissioners, providers and healthcare professionals.

“Pharmacy professionals should remain accountable for decisions within their professional control, but responsibility for risks, such as product defects, unsafe updates and organisational failures, should be allocated according to who is best placed to control and mitigate those risk across the AI lifecycle, rather than defaulting to frontline clinicians.

“We agree that clear, practical guidance is needed to help healthcare professionals understand their responsibilities when using AI, and we’re continuing to explore these issues with stakeholders through our digital and AI roundtables.”

Oputu added: “Recommendation 32 reflects a key point in our contribution to the [National] Commission’s work: pharmacy professionals need the right education, training and opportunities to develop the skills to use AI safely and effectively.

“Our existing policy also recognises the importance of building digital and AI capabilities across the pharmacy workforce, with education and training tailored to different roles and levels of interaction with AI.”

Kathie Cashell, chief executive of the General Pharmaceutical Council (GPhC), welcomed the report and said that the GPhC “will be carefully considering its recommendations”.

“We strongly agree with the recommendation that a coordinated approach is needed to building professional capability in the safe and effective use of AI in healthcare. We will continue to work closely with other regulators, the Professional Standards Authority and organisations across health and care to support safe, effective and responsible use of AI,” she continued.

“At the GPhC, we have recently published resources on the use of AI in pharmacy practice, revalidation and education and training and are planning further work in relation to AI as technologies continue to develop, including, for example, as part of the reviews of the standards we set for pharmacists and pharmacy technicians and our standards for registered pharmacies.”

Clarity on AI responsibilities in healthcare needed, National Commission says

Ferris State University is actively participating in Tech Week Grand Rapids, spotlighting the expanding influence of artificial intelligence (AI) across various sectors, particularly in healthcare.

Taking place from September 14 to 19, Tech Week Grand Rapids is a community-focused “unconference” comprising around 160 events throughout the Grand Rapids area, featuring panels, workshops, showcases from students, and networking sessions, all offered free to the public. Full event details can be found online.

Sessions during this week will delve into how AI, data, and emerging technologies are revolutionizing clinical practices, drug development, and patient care, with contributions from leading health tech organizations in the region.

Ferris State faculty and staff will host several noteworthy discussions, including:

**AI, Digital Equity, and Rural Education: Closing the AI Divide in Rural and Tribal Communities**  
Scheduled for Wednesday, September 16, from 11 to 11:45 a.m. at the Kendall College of Art and Design, this presentation will address the lack of focus on rural and Tribal areas in the ongoing AI revolution. As technological advancements predominantly benefit urban and suburban communities, the session will analyze the educational disparities faced by rural and Tribal populations in Michigan and propose solutions to bridge these divides.

**Leading AI Adoption: People Drive Success**  
On the same day from 1 to 2 p.m., this interactive session will explore the human dynamics of AI integration within organizations. It will highlight how leadership, effective communication, and change management are crucial for successful AI adoption—beyond just the technology itself. Attendees will engage in discussions and gain actionable insights applicable across industries.

**AI at the Bench: PKS Engineering Case Studies**  
Taking place on Thursday, September 17, from 11 a.m. to noon, this session will focus on the role of AI in engineering polyketide synthases (PKS), which are vital for producing chemicals used in biotechnology. Two student presentations will illustrate how AI can enhance the design-build-test-learn cycle in scientific inquiry, positioning AI as a collaborative tool rather than a standalone entity.

**CHARM: Practical AI for Healthcare Data**  
Also on September 17, from 1:15 to 2 p.m., attendees will learn about CHARM (Collaboration to Harmonize Antimicrobial Registry Measures), which utilizes healthcare data for improved antibiotic stewardship via interactive dashboards. This session will feature two AI-driven projects that streamline prescription data processing and leverage machine learning to detect outlier prescribing patterns, enhancing both accuracy and patient privacy.

**AI Adoption for Healthcare and Public Health: Practical Steps for Responsible Healthcare AI Adoption**  
That same afternoon, from 2:15 to 3 p.m., a presentation will discuss the transition from AI experimentation to effective implementation within healthcare systems. It will cover essential applications and address potential risks linked with privacy and bias, supported by a real-world case study illustrating how advanced data pipeline techniques can enhance antimicrobial stewardship.

**Applied LLMs in Pharmacy Practice: AI Projects Improving Pharmacy Workflows and Decisions**  
Concluding the day on September 17, from 3:15 to 4 p.m., this session will unveil innovative AI applications designed to support pharmacy practice. One project enhances access to data through a user-friendly conversational interface on a centralized dashboard, fostering faster and more efficient information retrieval while ensuring compliance with privacy standards. The second project automates various routine tasks in pharmacy service planning, significantly reducing manual workload and allowing professionals to concentrate more on patient care and strategic decisions.

Through these engaging sessions, Ferris State University aims to illuminate the pivotal role of AI in evolving healthcare and education, ultimately fostering a better understanding of its implications and potential in both urban and rural contexts.

Ferris State's Tech Week sessions in Grand Rapids emphasize the increasing importance of artificial intelligence.

A bright, sunlit conference hall at Ferris State University’s Grand Rapids campus, a large “Tech Week” banner draped across a warm wood-paneled stage, professional attendees in business casual clustered around demo stations. Sleek laptops display AI-generated data visualizations in natural tones, and a polished robotic arm gently hands a small chip to a researcher. Floor-to-ceiling windows reveal a Grand Rapids skyline under clear blue skies.

Ferris State’s Tech Week Grand Rapids sessions focus on the growing role of artificial intelligence - Ferris State University

- Ferris State University hosts Tech Week Grand Rapids, focusing on AI's societal impact.  
- Sessions address AI's role in healthcare, education, and addressing digital divides in communities.  
- Presentations showcase practical AI applications, improving workflows and decision-making in healthcare fields.  

Ferris Stateâs Tech Week Grand Rapids sessions focus on the growing role of artificial intelligence

GRAND RAPIDS, Mich. â

Ferris State University is playing a key role in Tech Week Grand Rapids presentations looking at how artificial intelligence is playing a growing role in health care and other aspects of society.

Tech Week Grand Rapids is a six-day, community driven âunconferenceâ taking place from Sept. 14 to 19 across the greater Grand Rapids area.

The event has about 160 events including panels, workshops, student showcases and networking mixers that are free and open to the public. Details about the entire event are located here.

These events will explore how AI, data, and emerging technology are transforming clinical care, drug discovery, and patient outcomes through a week of talks, demos, and hands-on sessions from the region's leading health tech organizations.

Sessions hosted by Ferris State faculty and staff are:

AI, Digital Equity, and Rural Education: Closing the AI Divide in Rural and Tribal Communities

Wednesday, Sept. 16, 11 to 11:45 a.m. at Ferris State Universityâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration information is here.

While society is in the middle of a burgeoning AI revolution, such unprecedented growth has largely been focused on urban and suburban spheres throughout the nation.

However, scant attention has been given to the unequal playing field in rural and Tribal communities, especially here, in Michigan.

This presentation will attempt to cover educational inequities and provide possible solutions to redress such trends.

The presentation will examine the growing trends in AI, but from an often overlooked and lesser-focused perspective, a rural and Tribal community-lens.

This session will not only document the current technological paradigm and disconnect between rural / Tribal and urban / suburban communities, but also note how this gap has been shrinking, as of late.

The session will also explore the paradox of AI in the rural and Tribal sectors, and show how it can be leveraged as a useful tool to bridge the proverbial 'digital divide,' but also how it risks further alienating such traditionally disenfranchised communities.

Leading AI Adoption: People Drive Success: Lead AI Change with Trust, Champions, and Confidence

Wednesday, Sept. 16, 1 to 2 p.m. at Ferris Stateâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration information is here.

AI is transforming every industry, but successful AI implementation is rarely limited by the technology itself. More often, organizations struggle because of uncertainty, resistance to change, unclear leadership, and a lack of organizational readiness.

This interactive presentation explores the human side of AI adoption and provides practical strategies for leading successful organizational change. Rather than focusing on the latest AI tools, we'll examine the leadership, communication, and change management practices that determine whether AI initiatives succeed or fail.

The session includes audience discussion, real-world examples, and actionable takeaways that leaders can immediately apply regardless of their industry or organization size. It is designed for business leaders, managers, HR professionals, educators, entrepreneurs, and anyone responsible for helping people navigate technological change.





AI at the Bench: PKS Engineering Case Studies

Thursday, Sept. 17, 11 a.m. to noon at Ferris Stateâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration information is here.

Polyketide synthases are large, assembly-line enzymes that manufacture a variety of chemicals essential for American biotechnological manufacturing, including fuels, medicines, and bulk chemicals. The engineering of PKS systems is complex, however, owing to the large size and complexity of the enzymes.

Engineering of these proteins requires repeated cycles of designing experiments, building genetic constructs, testing performance, and interpreting results. Artificial intelligence can support several stages of this designâbuildâtestâlearn, or DBTL, cycle when its outputs are converted into testable scientific hypotheses.

This session features two student-led presentations describing practical AI applications in PKS engineering. The speakers will demonstrate how AI can assist with hypothesis generation, experimental interpretation, identification of promising engineering strategies, and selection of subsequent experiments.

Together, the talks introduce the concept of âAI-in-the-DBTL-cycle,â placing AI within the scientific process rather than above it. AI serves as a scientific sounding board, helping undergraduates anticipate experimental outcomes and interpret their results more quickly.

CHARM: Practical AI for Healthcare Data: AI and Dashboards for Smarter Antibiotic Stewardship

Thursday, Sept. 17, 1:15 to 2 p.m. at Ferris State Universityâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration information is here.

How can AI move beyond experimentation and solve practical problems using real-world healthcare data?



This session will introduce Collaboration to Harmonize Antimicrobial Registry Measures â known as CHARM -- an antimicrobial stewardship initiative that transforms healthcare data into actionable insights through interactive dashboards and analytics.

Attendees will see how CHARM uses data from participating healthcare organizations to examine antibiotic prescribing patterns, identify opportunities for improvement, and support antimicrobial stewardship efforts.



The session will then showcase two AI projects being developed around the CHARM data ecosystem. The first uses locally hosted, open-source large language models to standardize prescription data while keeping sensitive healthcare data within a controlled environment.

The project combines structured data processing, retrieval-augmented generation, fine-tuning, systematic evaluation, and human review to improve accuracy while reducing hallucinations and manual data-cleaning effort.



The second project, CHARM Outlier, uses machine learning to identify antibiotic prescriptions that differ from expected prescribing patterns. Flagged records are presented through a web application where researchers and healthcare professionals can review them and provide feedback.

That feedback is then used to improve the modelâs future predictions, creating a human-in-the-loop approach to identifying potentially inappropriate prescribing.



Together, these projects demonstrate how dashboards, locally deployed LLMs, machine learning, and human expertise can work together to turn complex healthcare data into useful information while addressing practical concerns around privacy, accuracy, and responsible AI implementation.

AI Adoption for Healthcare and Public Health: Practical Steps for Responsible Healthcare AI Adoption

Thursday, Sept. 17, 2:15 to 3 p.m. at Ferris State Universityâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration is here.

Artificial intelligence is advancing rapidly, but many healthcare and public health organizations still struggle to move from experimentation to meaningful and sustainable adoption.



In this presentation and interactive Q&A, attendees will explore the difference between simply using an AI tool and successfully integrating AI into real organizational workflows.



Attendees will learn about practical AI applications such as data cleaning and standardization, automated reporting, prescribing outlier detection, dashboards, decision support, and natural-language analytics. The session will also examine important risks, including privacy, bias, hallucinations, incorrect outputs, explainability, and the need for human review.



A real-world case study from the CHARM will demonstrate how secure data pipelines, analytics, small language models, machine learning, and natural-language reporting can support antimicrobial stewardship.



The session is designed for healthcare professionals, public health leaders, technology professionals, students, researchers, business leaders, and organizations exploring practical AI implementation.

Applied LLMs in Pharmacy Practice: AI Projects Improving Pharmacy Workflows and Decisions

Thursday, Sept. 17, 3:15 to 4 p.m. at Ferris State Universityâs Kendall College of Art and Design, 17 Pearl Street NW in Grand Rapids.

Registration information is here.

We are presenting novel AI capabilities in support of the CHARM, whose centralized dashboard is hosted and supported by the Ferris State University College of Pharmacy.

The first of two completed projects extends that dashboard â a web-based Power BI environment â with a conversational interface. Instead of drilling through filters and report screens, users simply ask a question in plain language.

Behind the scenes, a text-to-SQL layer paired with a large language model interprets the question, translates it into a valid query, and returns the matching data.

Every request is validated before it runs, and privacy safeguards are built-in to keep patient data protected and HIPAA-compliant, so faster access never comes at the cost of trust.

âThe second project applies the same technology to the business side of pharmacy practice. It automates the groundwork behind pharmacy service planning: prospect identification, market analysis, and the assembly of a comprehensive report that a pharmacy owner or health-system leader can review and act on.

Work that once consumed many hours of manual research now takes a fraction of the time, freeing pharmacists and administrators to focus where their expertise actually matters â judgment, strategy, and patient care.

Ferris Stateâs Tech Week Grand Rapids sessions focus on the growing role of artificial intelligence

Regenstrief Institute and Suki, a pioneer in AI-driven ambient clinical documentation, are joining forces to investigate the implications of ambient AI technologies on healthcare practitioners, patients, and healthcare systems overall. Their collaboration aims to generate independent research and evidence that will enable healthcare organizations to grasp the effects of these technologies.

As ambient AI solutions become more woven into the fabric of healthcare workflows, health systems are confronted with significant inquiries regarding their influence on clinician documentation loads, communication between patients and providers, care quality, operational efficiency, and overall trust. Both Regenstrief and Suki are committed to promoting evidence-based strategies that empower healthcare organizations to make educated choices about the incorporation and utilization of these technologies.

Suki specializes in developing AI-driven ambient clinical intelligence tools that streamline documentation and administrative processes for clinicians. In contrast, Regenstrief Institute is globally recognized for its proficiency in health informatics, electronic health records, implementation science, and generating real-world evidence. Together, they aim to enrich the existing knowledge surrounding ambient AI and its applications within healthcare.

Rachel Patzer, PhD, MPH, who serves as the president and CEO of Regenstrief Institute, remarked on the necessity for healthcare organizations to rely on rigorous, independent evidence to comprehend the effectiveness of AI technologies and their real-world implications on clinicians, patients, and healthcare delivery. She stated, “As ambient AI continues to advance, research and evaluation will be essential in ensuring the responsible and effective implementation of these technologies.”

The buzz around ambient AI technologies stems from their promise to alleviate administrative strain, enhance clinical documentation, and enable clinicians to dedicate more time to patient interactions. Despite this potential, healthcare leaders, policymakers, and providers are still in search of concrete evidence regarding implementation success, outcomes, user satisfaction, and the long-term effects of these technologies. Ambient AI platforms aim to capture conversations between patients and clinicians while aiding in documentation, which may lead to reduced time spent on administrative duties, thereby fostering more attentive patient care.

As healthcare institutions assess the latest AI advancements, both Regenstrief and Suki stress the significance of transparency, accountability, and making decisions based on solid evidence. With its extensive background in healthcare data analytics, clinical informatics, and comprehensive evaluations of new technologies, Regenstrief is particularly well-equipped to address crucial questions about the real-world performance of AI tools in clinical settings.

Regenstrief collaborates with Suki to assess ambient AI in the healthcare sector.

A bright, modern hospital consultation room flooded with warm morning light. A physician sits at a sleek touchscreen console reviewing patient vitals while slim ambient AI sensors are discreetly mounted in the ceiling. On the screen side by side are the Regenstrief logo and the Suki logo, linked by a softly glowing waveform that weaves around medical charts and a stethoscope. Natural daylight streams through large windows, emphasizing a hopeful, innovative partnership in healthcare.

Regenstrief and Suki partner to evaluate ambient AI in healthcare - Regenstrief Institute

- Regenstrief Institute and Suki collaborate to study ambient AI's impact on healthcare delivery.  
- They aim to address clinician workload, communication quality, and operational efficiency concerns.  
- Research is essential for informed adoption of AI technologies in real-world healthcare settings.  

Organizations share a commitment to understanding how ambient AI technologies impact clinicians, patients and healthcare delivery

Regenstrief Institute and Suki, a leader in AI-powered ambient clinical documentation technology, are collaborating to explore how independent research and evidence generation can help healthcare organizations better understand the impact of ambient artificial intelligence technologies on clinicians, patients and healthcare systems.

As ambient AI tools become increasingly integrated into healthcare workflows, health systems face important questions about their effects on clinician documentation burden, patient-provider communication, care quality, operational efficiency and trust. Regenstrief and Suki share an interest in advancing evidence-based approaches that help healthcare organizations make informed decisions about the adoption and implementation of these technologies.

Suki develops AI-powered ambient clinical intelligence solutions that assist clinicians with documentation and administrative tasks. Regenstrief Institute is internationally recognized for its expertise in health informatics, electronic health records, implementation science and real-world evidence generation. Together, the organizations are exploring opportunities to contribute to a growing body of knowledge around ambient AI and its role in healthcare.

“Healthcare organizations need rigorous, independent evidence to understand not only whether AI technologies work, but how they affect clinicians, patients and care delivery in real-world settings,” said Rachel Patzer, PhD, MPH, president and CEO of Regenstrief Institute. “As ambient AI continues to evolve, research and evaluation will play a critical role in ensuring these technologies are implemented responsibly and effectively.”

Ambient AI technologies have gained attention for their potential to reduce administrative burden, support clinical documentation and allow clinicians to focus more fully on patient interactions. However, healthcare leaders, policymakers and providers continue to seek evidence regarding implementation, outcomes, user experience and long-term impact. Ambient AI platforms are designed to capture patient-clinician conversations and assist with documentation workflows, potentially reducing time spent on administrative tasks while supporting more focused patient care.

As healthcare organizations continue evaluating emerging AI technologies, both organizations emphasize the importance of transparency, accountability and evidence-based decision-making. Regenstrief’s longstanding expertise in healthcare data, clinical informatics and rigorous evaluations of innovative technologies positions it to help answer critical questions about how AI technologies perform in real-world clinical environments.

Regenstrief and Suki partner to evaluate ambient AI in healthcare

As autumn arrives, discussions surrounding cancer gain prominence in national dialogues. The interval between summer's end and winter's onset is marked by various awareness initiatives, inspiring survivor narratives, and crucial reminders to prioritize health check-ups. This year, these conversations are intermingling with a noteworthy trend: a growing number of Americans are seeking health information through artificial intelligence (AI).

A recent survey revealed that 76% of Gen Z adults and 63% of millennials are now relying on AI as their primary resource for healthcare inquiries, often before consulting a physician.

With just a few clicks, individuals can quickly search for symptoms, decipher medical jargon, and gain insights into potential health risks. When used judiciously, AI can serve as a valuable resource; however, simply obtaining information does not equate to knowing the appropriate next steps, particularly when it comes to cancer. Regular check-ups with healthcare professionals are essential for accurately detecting serious health issues, especially when symptoms may not seem relevant at first glance. This underscores the idea that while AI can assist, it should not lead our healthcare decisions.

In my experience as the president of a Fortune 500 company, I have learned that transforming ideas into actionable decisions requires careful guidance, as aimless actions can quickly lead to stagnation. Therefore, having a well-defined strategy is crucial to creating an effective healthcare plan.

Unfortunately, in the realm of personal health management, we continue to miss the mark.

Despite many years of awareness initiatives, nearly two-thirds of Americans still postpone essential health screenings. This reluctance is not due to ignorance about the importance of screenings; rather, life often gets in the way. Additionally, many individuals harbor fears about what they might discover or convince themselves that feeling well is enough to forgo check-ups.

Here, technology can play a pivotal role in mitigating these challenges. AI can help clarify various screening options or prompt individuals to prepare important questions for their medical appointments. However, the capabilities of AI have limitations.

AI cannot physically accompany you to an appointment if you're hesitant about going. It’s incapable of performing necessary procedures like mammograms and colonoscopies, nor can it analyze your entire health history to identify new potential conditions. Importantly, it cannot replace the reassuring voice of a loved one asking, “Have you gotten that checked?”

Supporting this idea, research on cancer screening behaviors indicates that many people view encouraging family members or friends to undergo screenings as an expression of love, which in turn influences their own health behaviors. This illustrates that the most effective health interventions often come from those who care about us, rather than through mere technological interfaces.

Like many families, my own has been affected by cancer. I have witnessed friends and loved ones face diagnoses, treatments, and the accompanying uncertainties. Such experiences fuel my commitment to encourage others to transition from uncertainty to action.

This principle also aligns with Aflac's mission: we support families who face challenges when screenings are overlooked or diagnoses are delayed, stepping in when health insurance falls short. Our perspective on the AI dialogue is rooted in real-world implications reflected in our claims data.

You can embody this principle by taking charge of your health. For that reason, AI should serve as a support tool rather than the primary guide in our healthcare journeys.

Virgil Miller holds the position of president at Aflac Incorporated and Aflac U.S., recognized as a leading provider of cancer insurance in both the United States and Japan.

The viewpoints expressed in this commentary are those of the author and do not necessarily represent the opinions of Fortune.

Aflac President: reasons AI shouldn't lead your healthcare decisions

A dimly lit hospital corridor under flickering red alarm lights, a sleek robotic arm holding a medical syringe looms over an empty gurney, while a determined doctor in scrubs steps forward, hand raised to block it. Above the double doors behind them, the Aflac logo shines boldly in stark white. Shadows stretch across cracked tile floors, conveying a sense of caution and unease about AI taking the wheel in healthcare.

Aflac President: why AI shouldn’t drive your healthcare - Fortune

- A significant number of Gen Z and millennials rely on AI for primary healthcare inquiries.  
- Many Americans delay health screenings despite understanding their importance, often due to life distractions.  
- AI should enhance, not replace, personal healthcare practices and the importance of human encouragement.  

Every fall, cancer starts sounding louder in the national conversation. The months between summer and winter bring multiple awareness campaigns, survivor stories, and reminders to get checked. That attention really matters. This year, however, that conversation is happening alongside another major shift: More Americans are turning to artificial intelligence (AI) for answers about their health.

A recent study found that 76% of Gen Z adults and 63% of millennials now turn to AI as their first line of primary healthcare — before ever seeing a doctor.

In seconds, we can search for symptoms, understand medical terminology and learn more about potential risks. Used responsibly, AI can be an extraordinarily useful tool, but knowing more is not the same as knowing what to do next. This is especially true when it comes to cancer. Consistent physician care can more accurately identify signs of more serious conditions associated with symptoms that might not always seem relevant. That is why AI should copilot but not drive our healthcare.

Being president of a Fortune 500 company means turning concepts into decisions, and decisions into action. After more than two decades in leadership roles at Aflac, I have learned that action without guidance can quickly turn into inaction. The best cure for that is a solid strategy that results in a viable care plan.

And this is where, when it comes to our own healthcare, we continue to fall short.

Despite decades of awareness campaigns, nearly two-thirds of Americans still delay recommended health screenings. It is not because people don’t understand that screenings matter; rather, life just gets busy. We also know that people worry about what they might learn, or convince themselves that because they feel fine, there is no reason to get checked.

Technology can help address some of those barriers. AI can be a component of the plan. It can help someone understand the difference between screening options or help prepare questions before a doctor’s appointment. But there is a limit to what AI technology can do.

AI cannot take you to an appointment when you are hemming and hawing about going. It cannot perform a mammogram or a colonoscopy. And it cannot personally view your health history and connect the dots on a possible condition that didn’t exist before, and certainly cannot replace the voice of someone you trust asking, “Did you get that checked?”

Separate research on cancer-screening behavior backs this up: most Americans describe encouraging a loved one to get screened as an act of love, and say that encouragement changes their own behavior. It’s a reminder that the most powerful health intervention tool is not information coming out of a screen. It often can be someone who knows and cares about you.

Like so many families, mine has been touched by cancer, and I have watched loved ones and friends navigate diagnoses, treatment and the uncertainty that comes with both. Those experiences are why I will continue to raise my voice to help move people from uncertainty to action.

That’s also, frankly, Aflac’s business model: we exist because a screening gets missed, a diagnosis comes late, and a family needs help covering what health insurance alone doesn’t. The AI conversation isn’t abstract for us — it shows up in our claims data.

You can apply this principle by being the leader of your health. That is why AI should be the copilot but not the driver of our healthcare.

Virgil Miller is president of Aflac Incorporated and Aflac U.S., a leading provider of cancer insurance in the United States and Japan.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

Aflac President: why AI shouldn’t drive your healthcare

Healthcare leaders must not mix up the potential of models with actual operational effectiveness.

The challenges faced within healthcare administration stem from a lack of cohesive information, fragmented workflows, and unclear accountability—not from an absence of data. For decades, the industry has invested heavily in systems designed to track various activities, such as electronic health records, billing software, payer portals, scheduling tools, call center systems, and analytical applications. While each of these systems captures vital information, very few are equipped to analyze the entire array of decisions that dictate whether patients receive timely care, if clinicians obtain the necessary documentation, and how providers are reimbursed correctly.

This is the challenge that artificial intelligence (AI) is now tasked with addressing.

The revenue cycle has emerged as a key area for AI testing in healthcare.

The revenue cycle encompasses the entire process that healthcare providers follow to secure payment for services rendered—initial scheduling and registration through coding, billing, payer interactions, and collection of payments.

This sector presents a unique opportunity for robust AI implementation due to its combination of high transaction volumes, intricate decision-making, and both structured and unstructured data, all of which necessitate measurable outcomes and exhibit widely varying operational dynamics. Moreover, the revenue cycle is pivotal, linking financial performance with patient access and administrative workload.

The complexity of a single insurance claim illustrates this intricacy; it can be impacted by factors such as the patient's insurance details, clinical documentation, coding regulations, payer policies, prior authorization prerequisites, medical necessity standards, and numerous other data inputs and operational facets. A failure in any of these areas can trigger significant repercussions weeks or even months later.

This complexity explains why one-size-fits-all automation has frequently been inadequate.

Traditional robotic process automation tends to excel in environments where workflows are stable and rules are clear-cut, but the reality of healthcare administration is quite the opposite. Payer demands shift. Documentation standards are in constant flux. Exceptions are prevalent and can be significantly impactful.

The next challenge for healthcare AI is integration.

A sunlit modern hospital operating room where a female surgeon in scrubs studies a sleek tablet displaying live patient vitals, diagnostic scans, and AI-suggested treatment paths. Warm natural light spills in through large windows, and soft, glowing circuit-pattern accents trace the edges of the tablet screen—symbolizing seamless integration of AI into healthcare.

Healthcare AI’s next test is integration - MIT Technology Review

- Healthcare's administrative challenges stem from fragmented information and workflows, not lack of data.  
- The revenue cycle process is highly suitable for rigorous AI deployment in healthcare.  
- Traditional automation fails due to unstable workflows and unpredictable payer requirements in healthcare.

But healthcare leaders should not confuse model capability with operational capability.

Healthcare’s administrative challenges are caused by fragmented information, fragmented workflows, and fragmented accountability, not a lack of information. The industry has spent decades investing in systems that capture activity: electronic health records, billing platforms, payer portals, scheduling systems, call center platforms, and analytics applications. Each system records something important. But few were designed to reason across the full chain of decisions that determines whether patients get timely access, clinicians have the right documentation and providers are reimbursed appropriately.

This is the problem that AI must now confront.

Revenue cycle is becoming one of healthcare AI’s proving grounds

The revenue cycle is the process healthcare providers use to get paid for care — from scheduling and registration through coding, billing, payer follow-up, and payment collection.

It is unusually suited to rigorous AI deployment because it combines high transaction volume, complex reasoning, structured and unstructured data, measurable outcomes, and significant operational variation. It also sits at the intersection of financial performance, patient access, and administrative workload.

A single claim can be influenced by patient insurance information, clinical documentation, coding rules, payer-specific policies, prior authorization requirements, medical necessity criteria, and many other data sources and operational processes. A breakdown in any one of those areas can create downstream consequences weeks or months later.

This is why generic automation has often fallen short.

Traditional robotic process automation works well when workflows are stable and rules are predictable, but healthcare administration is neither. Payer requirements change. Documentation expectations evolve. Exceptions are common and often material.

Artificial Intelligence in Healthcare Services in India

More from AI in healthcare

The lack of patient identity control is a gap in healthcare AI.

AI Healthcare Agents Speed Up Patient Access to Care

Bay Area healthcare system balances tech upgrades and patient trust while navigating AI.

Goldman selects Chinese healthcare stocks for a post-AI investment strategy.