Harvard Health Publishing extends a valuable service to our audience by granting access to a collection of archived materials. It's important to pay attention to the date indicated for the last review or update on each article.

Remember, the information available on this website, irrespective of its publication date, should never replace professional medical advice from your healthcare provider or any other certified medical expert.

AI in healthcare: Is a chatbot capable of responding to your medical inquiries?

A warm, sunlit modern living room: an adult sits on a light-gray sofa, holding a tablet that displays a medical chatbot interface labeled “AI Health Assistant.” Conversation bubbles read “My headache persists—what should I do?” and the chatbot’s reassuring response. On a nearby coffee table lie a stethoscope, prescription pad, and a cup of tea. A potted plant by the window adds a touch of calm, as soft natural light fills the scene.

AI in healthcare: Can a chatbot answer your medical questions? - Harvard Health

- Harvard Health Publishing offers access to a library of archived content for readers.  
- Articles include a date of last review or update for reference purposes.  
- Content on the site is not a substitute for professional medical advice.

As a service to our readers, Harvard Health Publishing provides access to our library of archived content. Please note the date of last review or update on all articles.

No content on this site, regardless of date, should ever be used as a substitute for direct medical advice from your doctor or other qualified clinician.

AI in healthcare: Can a chatbot answer your medical questions?

In Watertown, NY, the weather outlook for today features overcast skies accompanied by intermittent rainfall. The temperature is expected to peak at 72°F, with southwest winds blowing at speeds of 10 to 20 mph. There is an 80% likelihood of rain throughout the day.

As we move into the evening, expect continued rain, with temperatures dropping to a low of 56°F. Winds will shift to the south-southwest, maintaining a speed of 10 to 20 mph. The chance of rain increases to 90%, with total precipitation likely reaching close to half an inch.

Patients distrustful of governments and companies promoting AI as a solution for rural healthcare.

A dusky rural road at twilight, lined with cornfields fading into shadows. In the foreground, a small group of adult villagers in work clothes stand tensely around a sleek AI-powered telemedicine kiosk, its screen glowing softly but casting uneasy reflections on their worried faces. Behind them, a distant government building with a fluttering flag and a large roadside billboard reading “AI Health Solutions” loom under darkening skies. Natural, low golden light from the setting sun highlights the tension in the scene.

Patients wary of governments, companies pushing AI as a rural healthcare solution - nny360.com

- Today's weather: cloudy with rain, high temperature of 72°F.  
- Winds are shifting from southwest to south-southwest at 10-20 mph.  
- Tonight's low expected at 56°F with a 90% chance of rain.  

Watertown, NY (13601)

Today

Cloudy with periods of rain. High 72F. Winds SW at 10 to 20 mph. Chance of rain 80%..

Tonight

Periods of rain. Low 56F. Winds SSW at 10 to 20 mph. Chance of rain 90%. Rainfall near a half an inch.

Patients wary of governments, companies pushing AI as a rural healthcare solution

Artificial Intelligence & Machine Learning , Data Privacy , Data Security

Why Healthcare AI Vendor Risk Demands Stronger Oversight

Tom Walsh of tw-Security on Prioritizing High-Risk Vendors and AI Governance

Tom Walsh, founder and principal consultant, tw-Security

Healthcare organizations can't rely solely on vendors' assurances that artificial intelligence tools are secure and trustworthy. As AI adoption accelerates, providers need stronger governance, better oversight of third-party vendors and greater scrutiny of how AI affects patient data and clinical decisions, said Tom Walsh, founder and principal consultant at consultancy tw-Security.

See Also: A Darkening Landscape: AI, Friend and Foe of Cyber Resilience

Many healthcare organizations lack the staff and time to thoroughly assess hundreds of third-party vendors. Walsh recommends prioritizing attention on vendors - including AI technology firms - that pose the greatest risks based on their access to protected health information and personally identifiable information.

Along with those efforts, organizations also need AI governance frameworks, updated vendor assessments and greater transparency into how vendors develop, validate and oversee AI capabilities.

"I would ask for better documentation from the vendors to prove what human oversight they've implemented into their AI products and services," Walsh said. "We do not want something bad happening to a patient because we relied on an AI that wasn't accurate."

Walsh also warned that AI-powered clinical documentation tools - including ambient tools - can introduce errors into electronic health records if clinicians fail to review AI-generated notes before signing them. He advises healthcare organizations to strengthen human oversight, revise business associate agreements and update privacy notices so patients understand how AI collects, processes and uses their information.

"I don't think a lot of patients know or recognize that their conversation with a doctor or clinician is being scribed by an AI agent and then placed into their record," he said. "I think that's one area that patients are not aware of how AI is being used in hospitals or healthcare systems today."

In this video interview with ISMG, Walsh also discussed:

How to prioritize high-risk AI vendors for security and privacy reviews;

Risks related to shadow AI use in healthcare;

How AI-generated clinical documentation can affect data integrity and patient safety.

Walsh is founder, partner and principal consultant of tw-Security, a healthcare privacy and security consulting firm. He is a certified information systems security professional, a nationally recognized speaker and co-author of four books on healthcare information security. Walsh has over 34 years of information security experience. In addition to being an independent consultant, Walsh has been the virtual privacy and information security officer for several healthcare organizations and business associate firms.

Why Healthcare AI Vendor Risk Demands Stronger Oversight

Generate a cinematic news thumbnail for title:Why Healthcare AI Vendor Risk Demands Stronger Oversight. No text in image.

Why Healthcare AI Vendor Risk Demands Stronger Oversight - BankInfoSecurity

Until recently, we spoke of algorithms capable of identifying findings in medical images, prioritizing studies, or supporting physicians in diagnosis. Then came generative artificial intelligence, capable of summarizing information, drafting text, and creating content.

Now, a third stage is emerging: agentic AI. It does not merely generate a response. It can receive an objective, consult multiple sources, interpret context, execute a sequence of tasks, verify results, and escalate exceptions to a human.

In simple terms:

Traditional AI identifies or predicts.

Generative AI creates.

Agentic AI understands an objective, organizes actions, and executes a process.

Its impact on healthcare will not be limited to diagnosis. Agentic AI can also transform the vast universe of documentation surrounding each episode of care, including clinical notes, informed consent forms, referrals, laboratory results, radiology and pathology reports, nursing records, and clinical summaries.

An AI agent could receive documents from multiple sources, determine which patient they belong to, classify them, extract relevant information, identify missing data, integrate them into the appropriate medical record, and trigger alerts or follow-up tasks. It could also consult different systems to generate a preliminary draft of a clinical summary, referral, or discharge note, flag inconsistencies, and submit the result to the responsible healthcare professional for validation.

This is no longer simply about digitizing documents. It is about transforming dispersed content into contextualized, accessible, and actionable clinical information.

Autonomy Requires Governance

An AI agent operating without sufficient context can produce an answer that sounds convincing but is incorrect. If it works with fragmented, outdated, or poor-quality data, it may also automate an organization’s existing errors.

Agentic AI therefore requires interoperability and governed access to reliable information from electronic health records, PACS, RIS, laboratory systems, and document repositories. Its adoption must include clearly defined identities and permissions, traceability of every action, confidentiality safeguards, data provenance validation, human oversight proportional to risk, and mechanisms to stop, correct, or reverse automated actions.

The World Health Organization has emphasized that humans must remain in control of healthcare systems and medical decisions. In healthcare, even a seemingly administrative or documentation-related action can have clinical, legal, financial, and ethical consequences.

For Mexico and Latin America, where technological fragmentation, manual processes, and heavy administrative workloads remain significant challenges, agentic AI represents an opportunity to integrate information, reduce omissions, and give healthcare professionals more time to focus on patients.

However, the strategy should not begin by asking what artificial intelligence can do. It should begin by asking what problem it will solve, what information it will use, what actions it will be authorized to perform, and who will remain accountable for its results.

The next major transformation of AI in healthcare may not take place exclusively in front of a diagnostic image. It will also unfold across the millions of documents and data points accompanying every patient’s care journey.

Artificial intelligence can analyze, generate, and act. Responsibility, however, must remain human.

Norma Contreras is CEO of E+ Healthcare Technology Solutions and CEO of PaxeraHealth Mexico. An expert in healthcare IT, digital transformation, agentic AI, interoperability, cybersecurity, and health information governance, she has more than 25 years of experience leading strategic projects across Mexico and Latin America. She also serves as a business advisor to SMEs.

The Rise of Agentic AI in Healthcare Information Management

Generate a cinematic news thumbnail for title:The Rise of Agentic AI in Healthcare Information Management. No text in image.

The Rise of Agentic AI in Healthcare Information Management - Mexico Business News

Consumer-facing AI can improve public health—but more evidence is needed. Nat. Health 1, 159 (2026).

Plantin, J. -C., Lagoze, C., Edwards, P. N. & Sandvig, C. Infrastructure studies meet platform studies in the age of Google and Facebook. New Media Soc. 20, 293–310 (2018).

OpenAI. Introducing ChatGPT Health. https://openai.com/index/introducing-chatgpt-health/ (2026).

Staff, T. Perplexity health connects medical records and wearables in new AI push. TechRepublic https://www.techrepublic.com/article/news-perplexity-health-ai/ (2026).

Amazon. Introducing Health AI | Personalized insights, human-powered care. Amazon Health AI https://health.amazon.com/health-ai/learn-more

Microsoft. Copilot for Health: your AI Assistant. Microsoft Copilot https://www.microsoft.com/en-us/microsoft-copilot/for-individuals/do-more-with-ai/ai-for-daily-life/copilot-for-health-your-ai-companion-for-smarter-medical-decisions

Ant Group. Ant Group announces major upgrades to its 15-million-MAU AI health app AQ amid wider push into health sector. https://www.antgroup.com/en/news-media/press-releases/1765779300000 (2025).

South China Morning Post. Ant’s AQ health app tops 100 million users amid holiday surge. https://www.scmp.com/tech/article/3344359/ants-aq-health-app-breaks-through-100-million-users-amid-holiday-surge (2026).

Anthropic. Transform healthcare from insight to action. Healthcare | Claude by Anthropic https://claude.com/solutions/healthcare

Mandl, K. D. How AI could reshape health care—rise in direct-to-consumer models. JAMA 333, 1667–1669 (2025).

Lutes, M. E, & Hughes, R. IV. When the front door is an algorithm: guiding consumer use of AI in health care. Health Affairs Forefront https://doi.org/10.1377/forefront.20260302.877526 (2026).

Adler-Milstein, J., Murray, S. G. & Wachter, R. M. The market dynamics for third-party AI tools trying to compete with electronic health record developers. JAMA https://doi.org/10.1001/jama.2026.2394 (2026).

DataReportal. Digital 2026: Global Overview Report — Global Digital Insights. https://datareportal.com/reports/digital-2026-global-overview-report (2025).

Hwang, Y. -M., Ng, M. Y., Pillai, M., Sahai, M. P. & Hernandez-Boussard, T. The landscape of AI implementation in US hospitals. Nat. Health 1, 99–112 (2026).

Rutunda, S. et al. Large language models for frontline healthcare support in low-resource settings. Nat. Health 1, 191–197 (2026).

International Telecommunication Union. Measuring digital development: facts and figures 2025. https://www.itu.int/hub/publication/d-ind-ict_mdd-2025-3/ (2025).

Nimo, C. et al. AfriMed-QA: a pan-african, multi-specialty, medical question-answering benchmark dataset. in Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) (eds. Che, W. et al.) 1948–1973 https://doi.org/10.18653/v1/2025.acl-long.96 (Association for Computational Linguistics, 2025).

Ahia, O. et al. Do all languages cost the same? Tokenization in the era of commercial language models. in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (eds. Bouamor, H. et al.) 9904–9923 https://doi.org/10.18653/v1/2023.emnlp-main.614 (Association for Computational Linguistics, 2023).

Designated Health Professional Shortage Areas Statistics. Third Quarter of Fiscal Year 2026 Designated HPSA Quarterly Summary. https://data.hrsa.gov/Default/GenerateHPSAQuarterlyReport/?utm_source=openai (2026).

Journal of Urgent Care Medicine. One Medical: reimaging primary care around the consumer. https://www.jucm.com/one-medical-reimaging-primary-care-around-consumer/ (2016).

Mathis, W. S. et al. Vulnerability index approach to identify pharmacy deserts and keystone pharmacies. JAMA Netw. Open 8, e250715 (2025).

Catalano, G., Khan, M. M. M., Chatzipanagiotou, O. P. & Pawlik, T. Pharmacy accessibility and social vulnerability. JAMA Netw. Open 7, e2429755 (2024).

Fernández, E. V., Mcdaniel, J. & Carroll, N. Examination of the link between medication adherence and use of mail-order pharmacies in chronic disease states. J. Manag. Care Spec. Pharm. 22, 1247–1259 (2016).

Wang, X., Seyler, B. C., Han, W. & Pan, J. An integrated analysis of spatial access to the three-tier healthcare delivery system in China: a case study of Hainan Island. Int. J. Equity Health 20, 60 (2021).

Ye, P. et al. National-scale 1-km maps of hospital travel time and hospital accessibility in China. Sci. Data 11, 1130 (2024).

Du, X., Du, Y., Zhang, Y., Zhu, Y. & Yang, Y. Urban and rural disparities in general hospital accessibility within a Chinese metropolis. Sci. Rep. 14, 23359 (2024).

Huang, M. et al. Primary care quality and provider disparities in China: a standardized-patient-based study. Lancet Reg. Health West. Pac. 50, 101161 (2024).

Xu, J., Powell-Jackson, T. & Mills, A. Effectiveness of primary care gatekeeping: difference-in-differences evaluation of a pilot scheme in China. BMJ Glob. Health 5, e002792 (2020).

Huisman, L. et al. A digital mobile health platform increasing efficiency and transparency towards universal health coverage in low- and middle-income countries. Digit. Health 8, 20552076221092213 (2022).

Lucyk, K., Tang, K. & Quan, H. Barriers to data quality resulting from the process of coding health information to administrative data: a qualitative study. BMC Health Serv. Res. 17, 766 (2017).

Chernew, M. & Mintz, H. Administrative expenses in the US health care system: why so high?. JAMA 326, 1679–1680 (2021).

Yan, A. S., Apathy, N. C. & Chen, J. Adoption of health information technologies by area socioeconomic deprivation among US hospitals. JAMA Health Forum 6, e253035 (2025).

Adler-Milstein, J. & Jha, A. K. HITECH Act drove large gains in hospital electronic health record adoption. Health Affairs 36, 1416–1422 (2017).

Chishtie, J. et al. Use of epic electronic health record system for health care research: scoping review. J. Med. Int. Res. 25, e51003 (2023).

Wong, A. et al. External validation of a widely implemented proprietary sepsis prediction model in hospitalized patients. JAMA Intern. Med. 181, 1065–1070 (2021).

Capoot, A. Epic Systems is building more than 100 new AI features for doctors and patients. Here’s what’s coming. CNBC https://www.cnbc.com/2024/08/21/epic-systems-ugm-2024-ai-tools-in-mychart-cosmos-.html (2024).

Everson, J., Nong, P. & Richwine, C. Uptake of generative AI integrated with electronic health records in US hospitals. JAMA Netw. Open 8, e2549463 (2025).

Nong, P., Adler-Milstein, J., Apathy, N. C., Holmgren, A. J. & Everson, J. Current use and evaluation of artificial intelligence and predictive models in US hospitals. Health Affairs 44, 90–98 (2025).

Arora, R. K. et al. HealthBench: evaluating large language models towards improved human health. Preprint at https://doi.org/10.48550/arXiv.2505.08775 (2025).

Hicks, R. S. et al. HealthBench Professional: evaluating large language models on real clinician chats. Preprint at https://doi.org/10.48550/arXiv.2604.27470 (2026).

OpenAI. Making ChatGPT better for clinicians. https://openai.com/index/making-chatgpt-better-for-clinicians/ (2026).

Bulusu, P. Amazon launches Health AI agent on Amazon website and app with free 24/7 access to virtual care for Prime members. Amazon News https://www.aboutamazon.com/news/retail/amazon-health-ai-agent-one-medical/ (2026).

Businesswire. Ant Group launches major campaign against health misinformation via AI healthcare app AQ amid wider healthcare push. https://www.businesswire.com/news/home/20250818527127/en/Ant-Group-Launches-Major-Campaign-Against-Health-Misinformation-via-AI-Healthcare-App-AQ-Amid-Wider-Healthcare-Push/ (2025).

Anthropic. Advancing Claude in healthcare and the life sciences. https://www.anthropic.com/news/healthcare-life-sciences/ (2026).

Ramaswamy, A. et al. ChatGPT Health performance in a structured test of triage recommendations. Nat. Med. https://doi.org/10.1038/s41591-026-04297-7 (2026).

Center for Devices and Radiological Health. Clinical Decision Support Software. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-decision-support-software/ (2026).

Mello, M. M. & Cohen, I. G. Regulation of health and health care artificial intelligence. JAMA 333, 1769–1770 (2025).

Wang, M. Q. The ethics of Ant Afu: health AI within China’s super-app ecosystem. J. Med. Ethics https://blogs.bmj.com/medical-ethics/2026/01/30/the-ethics-of-ant-afu-health-ai-within-chinas-super-app-ecosystem/ (2026).

Sekalala, S. & Chatikobo, T. Colonialism in the new digital health agenda. BMJ Glob. Health 9, e014131 (2024).

Garg, S., Bebarta, K. K. & Tripathi, N. The Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (AB-PMJAY) after four years of implementation - is it making an impact on quality of inpatient care and financial protection in India?. BMC Health Serv. Res. 24, 919 (2024).

Belenguer, L. AI bias: exploring discriminatory algorithmic decision-making models and the application of possible machine-centric solutions adapted from the pharmaceutical industry. AI Ethics 2, 771–787 (2022).

World Health Organization. Ethics and governance of artificial intelligence for health. https://www.who.int/publications/i/item/9789240029200/ (2021).

Pan, J. et al. COMPASS-GH is a consensus roadmap for defining standards for safe, accurate and equitable AI in general health queries. Nat. Health 1, 162–163 (2026).

Glasgow, R. E., Vogt, T. M. & Boles, S. M. Evaluating the public health impact of health promotion interventions: the RE-AIM framework. Am. J. Public Health 89, 1322–1327 (1999).

Wang, H. et al. An evaluation framework for ambient digital scribing tools in clinical applications. NPJ Digit. Med. 8, 358 (2025).

Collins, G. S. et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ 385, e078378 (2024).

World Health Organization. Global strategy on digital health 2020–2025. https://www.who.int/publications/i/item/9789240020924/ (2021).

Greenhalgh, T. et al. Beyond adoption: a new framework for theorizing and evaluating nonadoption, abandonment, and challenges to the scale-up, spread, and sustainability of health and care technologies. J. Med. Int. Res. 19, e8775 (2017).

Torabi, F. et al. A common framework for health data governance standards. Nat. Med. 30, 26–29 (2024).

Shaw, J. & Sekalala, S. Health data justice: building new norms for health data governance. NPJ Digit. Med. 6, 30 (2023).

World Bank Group. Implementation know-how brief. Data governance in health. https://documents1.worldbank.org/curated/en/099081723223522777/pdf/P175075036726f0a00832307b7ae23a06a6.pdf (2025).

China Government Website. National Internet Information Office. Interim measures for the management of generative AI services. https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm (2023).

Omar, M. et al. Multi-model assurance analysis showing large language models are highly vulnerable to adversarial hallucination attacks during clinical decision support. Commun. Med. 5, 330 (2025).

Omiye, J. A., Lester, J. C., Spichak, S., Rotemberg, V. & Daneshjou, R. Large language models propagate race-based medicine. NPJ Digit. Med. 6, 195 (2023).

Obermeyer, Z., Powers, B., Vogeli, C. & Mullainathan, S. Dissecting racial bias in an algorithm used to manage the health of populations. Science 366, 447–453 (2019).

Wu, E. et al. How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals. Nat. Med. 27, 582–584 (2021).

Grundy, Q. et al. Data sharing practices of medicines related apps and the mobile ecosystem: traffic, content, and network analysis. BMJ https://doi.org/10.1136/bmj.l920 (2019).

Integration of consumer AI into healthcare pathways

Generate a cinematic news thumbnail for title:Integration of consumer AI into healthcare pathways. No text in image.

Integration of consumer AI into healthcare pathways - Nature

R1, a company that specializes in healthcare revenue cycle management, has announced its intention to acquire Humata Health, which focuses on utilizing artificial intelligence to streamline the prior authorization process. This move is aimed at enhancing efficiency in preapproval requirements that insurers impose for various medical treatments, thereby furthering R1's mission to automate payment processes for healthcare providers.

The acquisition is anticipated to be finalized in the third quarter of this year. Following the transaction, Humata’s team will work on integrating its innovative technology into R1’s existing revenue cycle platform, although specifics regarding the financial aspects of the deal were not disclosed by an R1 representative.

R1 is optimistic that this acquisition will bolster its AI capabilities, assisting healthcare providers in reducing administrative burdens and improving their financial outcomes. Operating out of Utah, R1 processes over 600 million transactions annually through a network of approximately 1,000 providers, utilizing a billing system designed to eliminate the necessity for multiple software applications.

In contrast to R1's diverse approach to claims processing, Humata Health, founded in 2023 by a physician trained at the Mayo Clinic, is dedicated solely to prior authorizations. Humata offers an AI-driven solution that recognizes insurer policies, generates prior authorization requests, and tracks them to final approval, boasting an impressive first-pass approval rate of up to 96%.

R1 CEO Joe Flanagan commented on the acquisition, stating, "Humata significantly enhances our coverage of the authorization process." The prior authorization system has long been a contentious issue within the healthcare sector, with patients expressing that the preapprovals create challenges, and providers pointing out the additional administrative tasks they entail, all while insurers argue that these authorizations are crucial for controlling costs and healthcare spending.

R1 envisions a future where real-time prior authorizations are commonplace, a vision echoed by recent federal regulations set to take effect in 2024 that will require insurers to expedite prior authorization decisions and use technology to facilitate electronic preapprovals. Additionally, during the previous summer, the Trump administration garnered commitments from insurers to ensure that at least 80% of electronic prior authorization requests receive real-time responses by next year.

Meanwhile, other technology firms are also launching solutions aimed at improving real-time prior authorizations. For instance, Epic, a leading electronic health record provider, recently introduced a tool that allows clinicians to quickly verify insurance prior authorization requirements.

R1 focuses on AI prior authorizations through acquisition of Humata.

A sunlit clinical office desk bathed in warm natural light, a heavy drill bit engraved with the R1 logo piercing through a neat stack of paper forms stamped “AI Prior Authorization,” and beside it a tablet screen glowing with the Humata logo; medical files and pens scattered around complete the scene of an acquisition driving deeper AI workflow integration.

R1 drills down on AI prior authorizations with Humata buy - Healthcare Dive

- R1 is acquiring Humata Health to enhance AI capabilities for prior authorizations.  
- The acquisition aims to streamline preapproval processes and reduce administrative burdens for providers.  
- Humata's technology boasts a 96% first-pass approval rate, improving authorization efficiency.  

Dive Brief:

Healthcare revenue cycle management company R1 has agreed to acquire Humata Health, a company specializing in artificial intelligence-powered prior authorizations.

R1 is banking that Humata’s AI expertise will help it streamline the preapproval process required by insurers for certain medical treatments, and further its goal of automating how healthcare providers get paid for their services.

The acquisition is expected to close in the third quarter this year, after which Humata’s team will start integrating its technology into R1’s revenue cycle platform. An R1 spokesperson declined to share the financial terms of the deal.

Dive Insight:

R1 says the acquisition should expand its AI capabilities while helping providers reduce paperwork and improve their bottom lines.

The Utah-based company, which processes over 600 million transactions annually across its network of 1,000 providers, says its billing platform eliminates the need for multiple disparate software tools.

While R1 has a broad focus on claims processing, Florida-based Humata Health is exclusively focused on prior authorizations.

The company, founded in 2023 by a Mayo Clinic-trained physician, offers an AI-powered product that identifies insurer policies, creates prior authorization requests and oversees them through final approval. Humata says its tool has up to a 96% first-pass approval rate on prior authorizations.

“Humata significantly enhances our coverage of the authorization process,” R1 CEO Joe Flanagan said in a statement.

The prior authorization process has long been a headache for the healthcare industry. Patients say the preapprovals are a burden, while providers say they increase administrative work and delay care. But insurers contend that prior authorizations are imperative to saving money and lowering healthcare spending.

R1 says its goal is to make real-time prior authorizations a reality. It’s a goal shared by the federal government — a rule finalized in 2024 requires insurers to speed up prior authorization decisions and implement tech to enable electronic preapprovals.

Last summer, the Trump administration secured voluntary commitments from insurers that at least 80% of electronic prior authorization approvals will be answered in real time next year.

Other IT companies are also debuting products they say help with real-time prior authorizations. Electronic health record company Epic launched a tool this week that enables clinicians to instantaneously check whether an insurer requires prior authorization.

R1 drills down on AI prior authorizations with Humata buy

Stay updated with the latest happenings in Minnesota through our informative newsletters. By subscribing, you'll receive not only major headlines and breaking news but also local stories, weather updates, and a variety of additional content. Our newsletters are sent directly to your inbox every weekday morning, ensuring you never miss out on essential news. Sign up today!

Mayo lawsuit underscores competition among hospitals to adopt AI, emphasizing the notion that "if you're not first, you're last."

A dimly lit hospital corridor at night, its walls cold concrete and flickering overhead lights. The Mayo Clinic logo glows boldly on a glass door at one end. Two rival teams of doctors in scrubs tug at thick, glowing cables leading into a central AI server chamber bathed in ominous red light. Above them, the shadow of a judge’s gavel looms large on the wall, casting tension over their frantic race to be first.

‘If you’re not first, you’re last’: Mayo lawsuit highlights tensions in hospitals' race to use AI - MPR News

- Minnesota newsletters offer headlines, breaking news, and local stories for subscribers.  
- Sign-up provides access to weather updates and essential information every weekday morning.  
- The service aims to enhance news consumption and keep residents informed regularly.  

News you can use in your inbox

When it comes to staying informed in Minnesota, our newsletters overdeliver. Sign-up now for headlines, breaking news, hometown stories, weather and much more. Delivered weekday mornings.

‘If you’re not first, you’re last’: Mayo lawsuit highlights tensions in hospitals' race to use AI

On Tuesday, September 8 at 9:00 a.m. ET, the Cyber Statecraft Initiative of the Atlantic Council will hold a public forum addressing the future of healthcare enhanced by artificial intelligence (AI) and the global conditions influencing its progress.

The landscape of healthcare is undergoing a significant transformation due to AI's impact on drug discovery, clinical trials, personalized medicine, and the overall delivery of healthcare services. However, to fully harness AI's capabilities, there is a critical need for access to high-caliber data, sufficient computing resources, reliable energy sources, robust digital infrastructure, and the ability to adapt to shifting regulatory landscapes. The discussion will focus on how international healthcare and pharmaceutical enterprises can effectively implement and scale AI technologies while maintaining patient trust, confidentiality, security, and sustainable practices.

Part of the Cyber Statecraft Initiative's Data Rules Project and backed by Pfizer, this conversation will feature key figures from the biosciences sector, government entities, and academia. They will explore how the US, China, and other leading technology hubs are tackling these pressing issues, and assess the implications of their varying policy decisions and infrastructural strategies on pharmaceutical innovation, patient outcomes, and the competitive landscape for the United States.

New policy updates: Attendees must present photo identification when arriving for the event, and on-site registration will not be available. Entrance will be closed twenty minutes after the event begins.

By securing a spot for in-person participation, attendees agree to adhere to the Atlantic Council’s Event Code of Conduct.

To register for this informative session, please fill out the form below.

Feel free to send questions for our speakers—your inquiry might be highlighted during the discussion.

Opening Remarks will be delivered by Jenna Ben-Yehuda, the Executive Vice President of the Atlantic Council.

Speakers include:
- Jeremy Forman, Chief AI, Data, & Analytics Officer at Pfizer
- Gil Alterovitz, Chief Artificial Intelligence Officer at the Department of Veterans Affairs and Associate Professor at Harvard Medical School

The session will be moderated by:
- Graham Brookie, Vice President of Atlantic Council Technology Programs
- Trey Herr, Assistant Professor at the School of International Service at American University and Nonresident Senior Fellow at the Cyber Statecraft Initiative, Atlantic Council

Additional speakers will be announced soon.

AI and healthcare data governance: From Washington to Beijing

A sunlit glass-walled conference room with a polished wood table. On one end sits a small U.S. flag, on the other a Chinese flag. Centered is a 3D-printed human brain model embedded with subtle glowing circuit tracings. Beside it lie a stethoscope and a tablet displaying patient data charts. Business delegates in sharp suits lean forward, pens poised over notepads, as regulatory maps and AI data flow diagrams are projected faintly on the glass walls—symbolizing the dialogue on AI healthcare governance from Washington to Beijing.

Data governance for AI and healthcare: From Washington to Beijing - Atlantic Council

- Atlantic Council hosts a discussion on AI-enabled healthcare and its global development challenges.  
- AI's potential in healthcare requires quality data, computing power, and regulatory navigation.  
- Leaders from biosciences and government will examine differing global policy approaches.  

On Tuesday, September 8 at 9:00 a.m. ET, the Atlantic Council’s Cyber Statecraft Initiative will host a public discussion on the future of AI-enabled healthcare and the global environment shaping its development.

AI is transforming drug discovery, clinical trials, precision medicine, and healthcare delivery. But realizing the full potential of AI in healthcare requires access to high-quality data, computing power, energy, and digital infrastructure, as well as the ability to navigate evolving regulatory requirements. How can global healthcare and pharmaceutical companies build and deploy AI at scale while protecting patient trust, privacy, security, and sustainability?

This discussion—part of the Cyber Statecraft Initiative’s Data Rules Project, supported by Pfizer—brings together leaders from the biosciences industry, government, and academia to examine the future of AI in healthcare. The discussion will also consider how the United States, China, and other major technology markets are approaching these challenges, and what their different policy and infrastructure choices mean for pharmaceutical research and development, patient outcomes, and US competitiveness.

New policy: In-person attendees will be required to show photo ID upon arrival, and no on-site registration will be permitted. Guests will not be admitted later than twenty minutes after the event’s start time.

By registering to attend in-person, you agree to abide by the Atlantic Council’s Event Code of Conduct.



Please complete the form below to register for this event.

Send our speakers a question—it may be featured during our event.





Welcome Remarks

Jenna Ben-Yehuda

Executive Vice President,

Atlantic Council

Speakers

Jeremy Forman

Chief AI, Data, & Analytics Officer,

Pfizer

Gil Alterovitz

Chief Artificial Intelligence Officer,

Department of Veterans Affairs;

Associate Professor,

Harvard Medical School



Moderated by

Graham Brookie

Vice President, Atlantic Council Technology Programs,

Atlantic Council

Trey Herr

Assistant Professor, School of International Service,

American University;

Nonresident Senior Fellow, Cyber Statecraft Initiative

Atlantic Council



More speakers to be announced.

Data governance for AI and healthcare: From Washington to Beijing

The integration of artificial intelligence (AI) into healthcare is expanding, focusing on risk prediction, patient monitoring, and automating administrative processes. A recent study from Oklahoma State University revealed that rural hospitals are falling behind urban counterparts in adopting AI, highlighting the need for investment in these underfunded facilities to safely evaluate and implement advanced technologies.

Brian Whitacre, a professor and extension economist at the university, aimed to explore the differences between urban and rural hospitals regarding their adoption, utilization, and assessment of predictive AI, as well as the factors influencing these rural institutions' willingness to integrate such technologies.

His research, currently undergoing peer review, leverages data from the 2023 American Hospital Association’s Information Technology Supplement, which for the first time included inquiries about the use of predictive AI models. Whitacre analyzed information from 1,301 urban and 1,018 rural non-federal acute care hospitals across the nation. The findings revealed a significant disparity in AI adoption, with rural hospitals utilizing predictive AI at a rate of 51%, compared to 81% for urban facilities.

“The urban-rural gap is substantial and should be acknowledged,” remarked Whitacre.

The motivations for adopting predictive AI vary considerably between rural and urban hospitals. The study found rural facilities are less likely to use AI for clinical decision-making or treatment recommendations, although they do use it for billing and operational efficiencies. Interestingly, rural hospitals predominantly rely on AI solutions developed by their electronic health record (EHR) vendors rather than creating their own or working with third-party developers, a trend contrasted by urban hospitals, which tend to utilize more self-developed or third-party AI models.

“Many EHR systems are shaping the AI adoption decisions,” Whitacre said.

In rural areas, AI is particularly beneficial for identifying high-risk outpatients needing follow-up care, an area where rural hospitals shine relative to their urban counterparts, as it directly impacts their revenue streams.

Several elements influence the probability of AI adoption in rural hospitals, such as their financial health and membership in larger health systems. Whitacre noted a positive correlation between stronger financial performance and the capacity to allocate funds towards productivity enhancements.

He cited Mercy Hospital Ada as an example, where officials are harnessing AI for various tasks, including compiling patient data, suggesting diagnoses, analyzing imaging tests, and documenting patient encounters. Dr. Benjamin Lynch, the hospital’s chief medical officer, emphasized that AI enables access to resources that independent rural hospitals often lack.

However, the findings also indicate that critical access hospitals and those owned by for-profit or governmental entities are less likely to adopt AI technologies, facing challenges like high initial costs and insufficient workforce capabilities for successful implementation.

Looking ahead, Whitacre sees his study as an initial overview. He believes more hospitals have likely adopted AI since the 2023 data collection, and he is interested in understanding the operational impacts of such technology on rural facilities.

“Will AI lead to increased revenue? Are patient outcomes improving?” Whitacre pondered. “There’s also the concern that if hospitals are utilizing AI efficiently, they might not hire as many staff, which isn't advantageous for rural communities.”

He is currently inviting collaboration with officials from rural Oklahoma hospitals and has consulted with leaders at Newman Memorial Hospital in Shattuck. CEO Tom Vasko, who serves on the advisory board for the Coalition for Health AI, noted that their partnership with the Mayo Clinic enhances cardiac remote patient monitoring for their patients. The AI-driven devices help foresee cardiac events.

Additionally, the implementation of DAX AI, which links to EHR systems to transcribe provider-patient conversations into clinical notes, improves interaction without compromising care quality. Vasko highlighted the importance of maintaining personal contact between patients and providers in rural settings, which fosters patient satisfaction and allows for increased patient throughput.

Avo, a clinical intelligence platform active in about 30 rural hospitals nationwide, is assessing how its AI tools are influencing financial outcomes for clients. Laurence Coman, Avo’s COO, pointed out how one critical access hospital in Montana has achieved zero claim denials since adopting their tools, which streamline billing processes and reduce administrative burdens.

While acknowledging apprehensions about AI potentially displacing jobs, Coman believes that these tools are vital for sustaining rural healthcare amid funding cuts. Avo has determined that its solutions for documentation and billing can help mitigate financial losses.

The adoption of AI in hospitals mirrors the initial integration of electronic health records. Early 2000s research showed rural hospitals were slow adopters compared to urban ones. Following that trend, the federal government initiated a $27 billion incentive program in 2009 under the Health Information Technology for Economic and Clinical Health Act to facilitate EHR adoption.

Whitacre argues for a similar initiative—dubbed HITECH 2.0—focused on promoting and evaluating AI applications in rural healthcare settings.

“This is going to be an ongoing issue. Early adopters will ultimately succeed, while those who hesitate may fall behind,” Whitacre concluded. “We need to assist these facilities in finding suitable case studies and support their AI integration efforts.”

AI adoption in healthcare is on the rise, but a researcher in Oklahoma discovered that rural hospitals are lagging behind.

A dimly lit farmhouse office at dusk, a lone researcher hunched over a laptop and spread-out maps of rural Oklahoma hospitals, warm desk lamp casting long shadows. Through the window, a modest country hospital sits under a muted streetlight, its windows dark. An Oklahoma state flag hangs on the wall.

AI use is growing in healthcare. An Oklahoma researcher found rural hospitals are slow to adopt it - KGOU

- Rural hospitals are adopting predictive AI at significantly lower rates than urban hospitals.  
- Financial health and system membership positively influence rural hospitals' AI adoption likelihood.  
- Challenges like costs and infrastructure hinder effective AI implementation in rural healthcare settings.  

Artificial intelligence is increasingly being integrated into healthcare to predict risks, monitor patients and automate administrative tasks. New research from Oklahoma State University found that rural hospitals are lagging in AI use and makes the case for investments in facilities that lack the funds to evaluate and implement it safely.

Brian Whitacre, a professor and extension economist in the university’s Department of Agricultural Economics, said he wanted to examine rural-urban differences in hospitals’ adoption, use and evaluation of predictive AI, and the factors that might make rural facilities more or less likely to adopt it.

His research paper, which is currently under peer review, uses responses from the 2023 American Hospital Association’s Information Technology Supplement. This year’s survey was the first to include questions about predictive AI model use.

Whitacre compiled submissions from 1,301 urban and 1,018 rural nonfederal acute care hospitals across the U.S. His research found that rural hospitals adopted and used predictive AI at substantially lower rates than urban hospitals, at 51% and 81%, respectively.

“The takeaway message from our paper is that the gap is real – the urban-rural gap is real,” Whitacre said.

Urban and rural hospitals adopt predictive AI for different reasons. Whitacre found that rural facilities are less likely than urban ones to use it for clinical decision-making or treatment recommendations, but equally likely to apply AI to billing and other operational tasks.

Rural hospitals are almost twice as likely to use models developed by their electronic health record provider than they are to create their own or hire a third-party developer. Urban hospitals are more likely to use third-party or self-developed models, but they still rely heavily on their electronic health record developers.

“A lot of the [electronic health records] are kind of driving the AI decision,” Whitacre said.

Rural hospitals are also more likely to use AI to identify high-risk outpatients for follow-up care.

“It makes sense that rural are actually outperforming urban on that basis, because that's where their bread and butter is,” Whitacre said. “They need that. That’s a good chunk of their revenue coming in.”

Certain factors were also linked to whether a rural hospital would adopt this technology. System membership and stronger financial health were positively associated with AI adoption.

Whitacre, for example, spoke with leadership at Mercy Hospital Ada, which is using AI to compile and share patient information, suggest possible diagnoses, analyze diagnostic imaging tests and record patient information. Dr. Benjamin Lynch, the hospital’s chief medical officer, said in a press release that they can leverage resources independent rural hospitals might lack.

Whitacre added that hospitals with higher operating margins are more likely to have extra revenue to devote toward enhancing productivity.

But the research found that critical access hospitals and facilities under for-profit or government ownership are linked to a lower likelihood of adoption.

Challenges to adoption include high upfront or per-use costs and lacking infrastructural and workforce capacity to ensure appropriate use.

The next step

The paper is just a snapshot, Whitacre said. More hospitals have likely adopted AI beyond the 2023 data he examined. His next question is, ‘Once rural hospitals do adopt AI, how does it impact their operations?’

“Does it mean they're having revenue increases? Are their patients having better outcomes?” Whitacre asked. “Or, I mean, the alternative is, are they using it super effectively, and they're not hiring as many people or even firing people? That's not really good for a rural community either.”

Whitacre currently has an open call out for rural Oklahoma hospital officials interested in collaborating on his research. He’s also spoken with leadership at Newman Memorial, a critical access hospital in Shattuck.

Jillian Taylor / StateImpact Oklahoma Newman Memorial Hospital in the northwestern Oklahoma town of Shattuck

CEO Tom Vasko is also on the advisory board for the Coalition for Health AI, whose mission is to advance the responsible development, deployment and oversight of AI in healthcare.

He said the facility in northwest Oklahoma is currently partnering with the Mayo Clinic to provide patients with advanced cardiac remote patient monitoring. Vasko said the devices come from the Mayo Clinic and are prescribed to Newman Memorial patients. The technology helps them predict cardiac events.

The hospital also recently implemented DAX AI, which connects to their electronic health record system. It can record conversations between a provider and patient, transforming them into a clinical note that providers can verify and sign off on. This creates more direct interaction time without damaging quality of care, Vasko said.

“In rural environments, it's so important to keep interaction face to face with the provider and the patient,” Vasko said. “The patients want to feel like they're being heard. And the providers, when you talk to them, one of the reasons why they love to practice out here is that they’re like, ‘I get to have a conversation with my patient versus being behind a keyboard and typing away as they're talking.’”

Some specialists come out to Shattuck only a couple of times a month, meaning they get booked out. Vasko said this tool allows providers to see more patients, improving access to care.

The program also helps with billing by ensuring providers meet payer requirements to avoid denials.

“It's helping the bottom lines of a lot of the rural hospitals,” Vasko said.

Vasko said the hospital hasn’t quantified the difference between operations now and before it implemented these tools.

Avo, a clinical intelligence platform serving around 30 rural hospitals across the U.S., is studying the return on investment its clients experience when utilizing its AI tools. The company offers resources like an ambient AI scribe integrated into an electronic health record system and a billing optimization tool.

Laurence Coman, the company's COO and co-founder, said a critical access hospital they serve in Montana has received zero denials on the claims they submitted since launching these offerings.

“There's so much inefficiency, and back and forth between the hospitals, and then insurance,” Coman said. “[It removes] such a big headache from a critical access hospital that might have one biller, and that biller might be also doing IT work or other administrative work.”

He said he understands there are fears associated with AI’s potential to replace jobs. But he believes tools like this can help support the people working in and the finances of rural healthcare – especially amid cuts to federal Medicaid spending . Avo has calculated that AI tools for documentation and billing can help offset some of these looming losses.

“I think that's really important framing for why implementing doesn't just have to be Avo and my company's technology, but why implementing AI is more critical than ever – or really any technology that can help them with these problems,” Coman said. “It's just like a very critical juncture for rural hospitals.”

Incentivizing rural hospitals

Whitacre said the diffusion of AI into hospitals is similar to the spread of electronic health records. Research in the early 2000s documented lower adoption among rural settings compared to urban counterparts.

In 2009, the federal government set aside $27 billion through the Health Information Technology for Economic and Clinical Health Act for an incentive program to encourage hospitals and providers to adopt electronic health record systems. Billions more were allocated to help train workers and assist facilities in setting up these systems.

Whitacre argues there should be a HITECH 2.0 to support rural hospitals in adopting and evaluating AI tools.

“It's not going away. It seems like there's going to be winners, and probably the losers are going to be people that don't adopt it early,” Whitacre said. “So maybe we need to help these laggards get it up and running – find the right case studies for their hospitals. So that's what the paper hopefully makes a case for.”

AI use is growing in healthcare. An Oklahoma researcher found rural hospitals are slow to adopt it

For the past two years, 83-year-old Keith Magnuson has been grappling with severe pain due to lumbar spinal stenosis, a debilitating back condition that makes even simple tasks unbearable. Previously an avid rock climber and hiker, he now struggles to unload his dishwasher without discomfort. 

"I’m in pain every time I’m on my feet," Magnuson, a resident of Seattle, lamented. Thankfully, his doctor proposed a minimally invasive procedure called lumbar decompression (MILD), which poses a lower risk than traditional surgery and is covered by Medicare for patients over 65. However, Magnuson faced a hurdle: an epidural steroid injection, part of his treatment plan, was denied by Medicare.

Despite his doctors submitting additional requests with further documentation, the denial persisted. It was only during the third rejection that he learned the decision was made not by a medical professional, but by artificial intelligence. “I was outraged,” Magnuson remarked. “It’s AI making these decisions?”

Magnuson's situation isn't an isolated incident. Across six states in the U.S.—Washington, Arizona, New Jersey, Ohio, Oklahoma, and Texas—Medicare enrollees are discovering that AI programs are determining their eligibility for specific treatments. This pilot program, initiated by the U.S. Department of Health and Human Services (HHS), aims to assess whether AI can curb excessive spending within Medicare by denying “unnecessary” medical procedures.

Medicare, a public health insurance program mainly for seniors and some individuals with disabilities, is a significant component of federal spending, accounting for about 14 percent. The program has initiated this AI endeavor as part of a broader initiative to minimize costs and prevent inappropriate treatment recommendations, especially after a report suggested that up to $5.8 billion in Medicare spending in 2022 was deemed wasteful.

This pilot program, known as the Wasteful and Inappropriate Service Reduction (WISeR) Model, requires healthcare providers to submit justifications online for particular treatments. A third-party AI assesses these submissions to decide whether they meet Medicare criteria.

Magnuson represents just one case among many, as the implementation of the WISeR initiative has led to significant complications. Medical providers are experiencing technical difficulties, prolonged wait times, and unexpected treatment denials. Hospitals often find themselves inundated with additional paperwork and the need to bring patients in for more appointments. Approval for procedures has transitioned from a matter of days to potentially weeks, leaving patients like Magnuson in severe discomfort.

"This was presented to us as something that would revolutionize the process, ensuring swift approvals," explained Jeb Shepard, director of policy at the Washington State Medical Association. In reality, however, patients are frequently waiting weeks, a situation especially detrimental for the elderly, whose health can deteriorate rapidly without prompt care.

Despite some tech companies offering timely assistance regarding care denials, others have proven to be unresponsive, leaving physicians frustrated. Dr. Jeff Marr, a health economist, described the AI decision-making process as a "complete black box," highlighting the lack of transparency around how these systems operate.

The WISeR program scrutinizes 15 types of treatments, including epidural steroid injections, and necessitates prior authorization from the AI each time a doctor prescribes one. While the Centers for Medicare and Medicaid Services (CMS) claims doctors should receive results within 72 hours, the reality has been far more tangled. Health professionals complain that this added layer of AI oversight complicates an already simplified approval process, turning it cumbersome.

Critics argue that AI should not dictate treatment approvals, with Dr. Steve Aydin, a pain specialist in New Jersey, asserting that medical decisions require a clinician’s nuanced understanding of a patient’s individual experience. Many physicians express concerns that the deployment of AI lacked adequate consultation, leading to administrative obstacles and patient delays.

Michelle Mello, a professor of health policy, notes that this program was rolled out without Congressional approval, allowing the administration to bypass a formal procedure due to its voluntary nature for tech companies. Nevertheless, the program is mandatory for doctors, who risk losing Medicare funding without participation.

As the effects of the WISeR program become apparent, many have expressed their worries that the payment structures for AI services could inadvertently encourage unnecessary care denials. Frustrations mount as patients consistently face delays in receiving much-needed medical attention.

In this context, individuals like Magnuson have taken measures into their own hands, sometimes opting to pay for treatments out of pocket, even when it leads to uncertainty about reimbursement. Magnuson’s plea highlights the systemic issues within the rollout of AI in Medicare, urging for a reconsideration of how such technology is implemented in healthcare.

While many within the medical community acknowledge the potential benefits of AI—such as improving diagnostic accuracy—there is a pervasive concern about entrusting crucial decision-making solely to machines, devoid of human judgment. The challenge ahead will be ensuring AI serves as a complement to human expertise rather than a replacement, particularly when patient lives are at stake.

AI has started to determine healthcare access in six US states, and it is already negatively impacting patients.

A dimly lit hospital corridor bathed in deep red emergency lights, a robotic arm holding a patient file stamped “DENIED,” six U.S. state outlines glowing faintly on a wall map behind it, anxious adult silhouettes waiting under harsh shadows.

AI is now deciding who gets healthcare in 6 US states – and it's already harming patients - BBC Science Focus Magazine

- 83-year-old Keith Magnuson faced treatment denials due to Medicare's AI-driven approval system.  
- Medicare's pilot program using AI has created delays and increased burdens for patients and doctors.  
- Concerns rise over AI decision-making replacing human judgment in healthcare treatment approvals.

Over the past two years, Keith Magnuson has been in debilitating pain. The 83-year-old has lumbar spinal stenosis, a back condition that causes problems any time he is on his feet.

Until recently, Magnuson was a regular rock climber and hiker, but these days it hurts just to unload the dishwasher.

“Any time I’m on my feet, I’m in pain,” says Magnuson, who lives in Seattle.

His doctor had found a solution. Magnuson could have a procedure called minimally invasive lumbar decompression (MILD), which would help his back and carries less risk than surgery.

As Magnuson was over 65, the procedure would be covered by his Medicare insurance and could be arranged and performed quickly.

There was just one problem. One part of the planned procedure – an epidural steroid injection to help address the back pain – had been denied by Medicare.

Magnuson’s doctors asked for approval again after providing more documentation, and then a third time. On each occasion, it was denied.

It was on the third denial that Magnuson discovered it wasn’t a person who had made this decision about his treatment. It was AI.

“I was outraged,” Magnuson says. “I was like, wait a minute, it’s not even another person at the other end? It’s AI? It’s a bot?”

A pilot with problems

Magnuson is not alone. Across the US, in six specially selected states – Washington, Arizona, New Jersey, Ohio, Oklahoma and Texas – people enrolled in Medicare are finding that AI is deciding whether or not they can receive certain treatments.

These states have been involuntarily enrolled in a six-year pilot programme by the US Department of Health and Human Services (HHS), which wants to find out whether AI can protect US taxpayers from “wasteful” spending within Medicare, and prevent patients from receiving “unnecessary” procedures.

Medicare is a government-run health insurance programme for the over-65s and some younger people with disabilities.

There isn’t an equivalent of the UK’s National Health Service in the US, so anyone not covered by Medicare has to pay for their healthcare either through employer-sponsored insurance or out of their own pockets.

Today, Medicare accounts for around 14 per cent of US federal spending.

In a bid to cut costs, the HHS wants AI to act as a safeguard against doctors prescribing treatments that patients may not need, citing a report from the Medicare Payment Advisory Commission that estimated up to $5.8bn spent on Medicare in 2022 was on “services with minimal benefit”.

The pilot programme forms part of the Trump administration’s embrace of AI as part of its efforts to achieve “global dominance” in the sector, which includes reducing the barriers to the use of AI within healthcare.

Under the pilot scheme – called the Wasteful and Inappropriate Service Reduction (WISeR) Model – certain procedures now require doctors to upload their reasoning for why a treatment is necessary to an online portal.

AI then assesses whether it can go ahead under Medicare cover.

83-year-old Keith Magnuson has been denied back-pain treatment three times since January - Image credit: Ken Lambert/Seattle Times

The AI is provided by third-party tech companies selected by the Centers for Medicare and Medicaid Services (CMS), which oversees the scheme, with each state assigned a different tech firm.

But hopes of using AI to cut down on waste and fraud, as well as make life easier for doctors, are already being dashed.

The WISeR scheme has got off to a rocky start, with mixed results across the country. Doctors have reported delays, technical glitches and unexpected denials of care.

Hospitals and medical practices trying to resolve these problems face extra paperwork and often have to call patients back in for more appointments.

Treatments that once took a day to authorise are now taking weeks. In the process, patients are being left in sometimes unbearable pain.

“This was sold to the physician community as: ‘We’re going to turn this around in 72 hours, it’s super fast’,” says Jeb Shepard, director of policy at the Washington State Medical Association (WSMA), “and instead we’ve had people waiting four weeks or beyond.

"And this is for a patient population that’s elderly – if they don’t get timely care, their conditions deteriorate pretty rapidly compared to someone who is middle-aged or younger.”

In addition, while some of the tech companies are responsive in explaining why care has been denied, others have been hard to reach by physicians, or have struggled to explain why a treatment has not been approved.

“I think it’s a complete black box,” says Dr Jeff Marr, a health economist and assistant professor of health services, policy and practice at Brown University’s School of Public Health.

“How these models work, what information they’re looking at, what data they’re trained on. I think there is very little that is able to be understood by researchers or the public.”

How WISeR works

There are 15 treatments covered by the WISeR scheme, ranging from epidural steroid injections for pain management to skin substitutes and knee arthroscopy.

The CMS says these have been selected because they are either often overused or have historically had a greater risk of fraud, waste or abuse.

Doctors must get permission – known as prior authorisation – from the AI each time they want to use one of these treatments.

The CMS says doctors should receive approval or denial within 72 hours. If the AI denies a treatment, a human clinician at the tech company must then manually review that decision.

The use of prior authorisation has frustrated both doctors and patients. Traditional Medicare has historically required prior authorisation only in limited circumstances, unlike private insurance plans where it is common practice.

Physicians argue that the government has inserted AI into a system that was already relatively seamless, and created a cumbersome process in its place.

This argument forms part of a wider complaint from physicians, health bodies and local politicians that there was very little consultation before the pilot programme was rolled out, despite growing concerns about using AI in healthcare decisions.

“I don’t think AI has a place in making approvals on whether a treatment is appropriate or not,” says Dr Steve Aydin, a pain doctor at Kayal Orthopaedic Center in New Jersey.

“A clinician is making a decision based on information that’s synthesised from an evaluation, a conversation, a history and a physical exam – the AI doesn’t know what the patient is feeling or going through.”

Shepard says there was “very little opportunity to share concerns about the potential administrative challenges and patient care delays. It was sort of like the administration said: ‘We’re doing this, and you’re subject to it’.”

Michelle Mello, a professor of health policy at Stanford University School of Medicine and professor of law at Stanford Law School, says the Trump administration was able to implement the scheme without Congressional approval because it has been classed as voluntary, meaning it does not have to go through a formal rulemaking process.

However, it is voluntary only for the tech companies providing the AI tools for this scheme. Doctors who do not participate do not receive Medicare funding

The new WISeR scheme has limited and slowed approval processes for many Medicare procedures - Image credit: Alamy

Mello adds that there are very few specific laws around the use of AI in US healthcare.

Parts of AI are regulated under the US Food and Drug Administration (FDA), but the relevant statute dates from 1976 – well before the advent of advanced computing, let alone AI.

“The FDA has, over the years, issued opinions stating that it thinks certain kinds of software can be considered a medical device, and have to go through FDA review,” she says.

“But in practice, around 95 per cent of what it has reviewed are algorithms that work in radiological devices and other physical things.

"So the kind of AI that we tend to worry about are not those things, they’re just freestanding algorithms. In theory, the FDA regulates some of those too, but it’s very circumscribed, and under the Trump administration, it’s become more so.”

Read more:

Delays and denials

Naturally, the administration disagrees, with Dr Mehmet Oz, the CMS administrator, saying: “Combining the speed of technology and the experienced clinicians, this new model helps bring Medicare into the 21st century.”

Those who have used the system say the reality is very different from Oz’s vision. In Washington state, where Magnuson lives, the rollout has been plagued with problems.

Physicians and patients have reported significant delays and unexpected denials from the selected tech firm, Virtix.

Maria Cantwell, a Democratic senator for Washington who is calling for WISeR to be scrapped, issued a report saying that Medicare patients in the state are waiting two to four times longer to complete procedures covered by the pilot programme, with an average wait of 15 to 20 days.

At the Northwest Endovascular Surgery in Richland, Washington, the use of AI in prior authorisation initially meant waiting up to a month for a result, though this has since been reduced to about a week, they say.

For Yasamin Alazawi, an intern and scrub technician at the surgery who is responsible for submitting prior authorisations, such delays are difficult to explain to patients.

“The patients that we see, they’re elderly. They’re in debilitating 10-out-of-10 pain.

"So having to explain to them that, ever since January, with WISeR you need to wait for however long it takes to hear back from Medicare and Virtix to schedule them – you can just hear that disappointment in their voice. It’s heartbreaking.”

In New Jersey, Aydin says unexpected denials are causing problems too, and have forced him to recruit more staff. Denials also mean calling patients back in for further appointments to submit a new prior authorisation request with additional documentation.

“We’ve gone from a scenario where we could have just treated a patient right away, even sometimes on the same day, and turned it into potentially three separate visits,” Aydin says.

Physicians have also reported difficulties in communicating with tech firms to find out why a case is delayed or has been denied.

In Washington, Alazawi says of Virtix: “In the first month, I couldn’t even get anyone on the phone. I would be on hold forever, and I just wouldn’t get anywhere.

“That has improved, but it still doesn’t excuse that first one to two months of just not being able to get a hold of anyone.”

Virtix did not respond to multiple requests for comment from BBC Science Focus but told The Seattle Times in March that delayed responses were caused by an unexpectedly high volume of prior authorisation requests.

The CMS has said it is monitoring turnaround times across all six states and that delays will result in “corrective action”. It adds that participants with a high rate of inaccuracy may be terminated from the model.

Virtix said that denials may occur when a prior authorisation submission lacks certain documents, but that doctors can resubmit requests or arrange a peer-to-peer review to discuss the denial with a Virtix doctor.

Alazawi says she is sceptical of this reasoning.

“In the denials that I have personally seen, it almost sounds like nobody checked the supporting documentation to actually see what we wrote, because their reasoning for denial usually doesn’t match up with the findings of the doctor. Every single appeal that we have submitted has been overturned.”

Hospitals and patients have also raised concerns about the payment structure of the WISeR model.

Patients often wait weeks, needing multiple visits to the doctors before treatment approval - Image credit: Getty Images

The amount tech companies are paid is partly based on the savings they generate from denying treatments, leading to fears of a perverse incentive to tweak AI algorithms to deny a certain number of treatments.

“You’re incentivising an entity to deny care,” Shepard says.

“The CMS would disagree with the way that we feel about this, but in a scenario where you tell a company with a profit motive that they get a cut of whatever inappropriate care is denied, it is hard to imagine that this doesn’t lead to more denials.”

The CMS argues that payments to tech firms are adjusted based on metrics such as customer satisfaction and how often a company’s initial decision on prior authorisation is overturned, meaning it is against a company’s interest to incorrectly deny care.

The case for AI

Those providing the AI tools say the technology has the potential to transform how doctors do their jobs.

Jeremy Friese, chief executive of Humata Health – which is providing the AI for prior authorisation in Oklahoma – is a former physician who knows how frustrating prior authorisation can be.

“Our reason for existing is to help get the yes [in prior authorisation requests], as efficiently and as fast as possible,” he says. “While prior authorisation can be a very friction-laden, problematic process, if you do it with technology, it can be much more streamlined.”

Friese says that after doctors upload clinical information about the procedure they want to perform, Humata’s AI checks that request against existing CMS regulations to see whether the information provided meets the agency’s requirements.

“If the answer is yes, we give them an immediate answer so they can immediately go deliver care,” he adds. “If the AI says: ‘I’m not sure that I see all the information’, then there is always a nurse or a doctor that reviews that clinical information.

"What I can tell you with absolute certainty is our artificial intelligence cannot be used to say no. It can only be used to say yes.”

Friese says Humata’s AI tool already has an 88-per-cent accuracy rate, and is constantly updating and learning from errors as it goes.

“No software is 100-per-cent accurate but we fully expect this will get into the high 90s over the coming months.”

A question of trust

For Magnuson, the delays and denials around his care eventually led him to offer to pay for the epidural himself, at a potential cost of $600. Since receiving the operation, he’s still waiting to find out whether he has to pay.

“What I’m trying to do [by speaking out] is help so many people who cannot come up with an extra $600 for this and should not have to,” he says.

He believes the Trump administration rolled out the WISeR programme without sufficient care or thought.

“They mindlessly decided to do this. They didn’t go to the public, they didn’t even go to Congress to get it through some sort of law or rule. They just decided to do it. And that’s a pretty bad way to do business.”

Many physicians involved in the WISeR scheme acknowledge that AI has its benefits, too. Computer vision will enable AI to detect problems in scans that doctors may miss, and its vast corpus of knowledge could help identify new potential treatments for diseases.

And the rollout of AI is not limited to Medicare, or even to the US.

In the UK, the tech firm Palantir has signed a £330m ($450m) deal with the NHS to help organise data held by NHS hospital trusts – currently in separate databases – into a single platform.

In Australia, BreastScreen NSW is using machine-reading technology to help radiologists interpret mammography images.

A study from Sweden found that using AI in breast cancer screening reduced the number of cancers diagnosed in the months after a routine appointment by 12 per cent – because the AI was catching more of them at the screening stage itself.

However, these are mostly scenarios in which humans lead the decision-making. What worries doctors is when humans are removed from the equation entirely.

“Do I think AI is great in the sense of, say, getting a diagnosis? Yes,” Aydin says. “Do I think it’s going to replace the true decision-making of a good and experienced doctor? I don’t think so.”

Read more:

AI is now deciding who gets healthcare in 6 US states – and it's already harming patients

The healthcare sector has a rich history of embracing innovation, continually advancing through technologies like electronic health records, connected medical devices, robotic surgery, sophisticated imaging techniques, and the integration of artificial intelligence. However, unlike other industries, the adoption of new technologies in healthcare hinges not merely on feasibility but on the level of trust they command.

The dialogue surrounding AI in healthcare is evolving. Instead of merely asking, "Can AI assist clinicians effectively?" the focus has shifted to, "Can we trust AI sufficiently to integrate it into clinical workflows?"

Establishing this trust requires more than just a single application or model; it must be cultivated across every facet of the technological framework—from the data that healthcare professionals rely on to the security measures safeguarding sensitive information, as well as the workflows that ultimately impact patient care. This is fundamentally the vision behind Cisco's Compatible Solutions for AI. By uniting Cisco's robust infrastructure with a growing network of verified software partners, we’re empowering healthcare providers to adopt AI with greater assurance while fostering trust throughout each component of the solution.

Trust Starts with Enhanced Clinical Insights

Hospitals generate a staggering amount of physiological data every second, gathered from monitors, ventilators, and infusion pumps, among other bedside devices. The challenge lies in consolidating this data in a manner that enables clinicians to access key information for evaluating patient conditions.

This is where Sickbay stands out. Utilizing Cisco Unified Edge, Sickbay aggregates high-frequency patient monitoring data from various bedside devices into one coherent interface. Rather than switching between numerous systems, clinicians can access a unified dashboard that presents a comprehensive view, complete with proactive alerts that streamline data visibility for healthcare teams.

By combining continuous waveform data with clinical information and electronic health records (EHR), AI can help spot patterns for clinical evaluation, diminishing alarm fatigue, facilitating timely access to data, and enhancing clinicians' confidence in their decisions.

Trust Requires Protection of Sensitive Data

As AI becomes integrated into healthcare processes, organizations face the critical challenge of securing some of the most private information while pursuing innovation. While healthcare has always demanded robust security measures, the advent of AI brings additional complexities. Protecting more than just patient records is crucial; models, prompts, inference data, and intellectual property also become vulnerable.

In this context, our collaboration with Fortanix takes a significant step forward. Together, we’re empowering organizations to operate AI workloads in confidential computing environments, where both models and sensitive data remain secure even during active processing. By employing hardware-based isolation, attestation, and zero-trust strategies, we enable organizations to move forward with AI on sensitive data without unwarranted risks.

For the healthcare sector, trust encompasses more than mere data encryption. It involves creating a secure infrastructure capable of supporting AI operations while handling sensitive information.

Trust Encompasses Everyday Care

Significant elements of clinical care and recovery frequently take place beyond traditional clinical settings. Consider something as fundamental as meal delivery, which plays a critical role in patient care by ensuring that dietary guidelines are followed.

Proppos employs computer vision and AI technology to confirm that a patient's meal aligns with their dietary restrictions prior to leaving the kitchen. While seemingly straightforward, this application highlights a much broader principle.

Healthcare AI serves not only to facilitate complex clinical evaluations but also to minimize preventable errors, enhance consistency, and enable clinical teams to concentrate on core tasks rather than manual checks.

Confidence is Key to Healthcare AI

Despite addressing vastly different challenges, these examples lead to a common realization: healthcare organizations will evaluate the success of AI based on the trust and confidence those solutions instill in clinicians, IT professionals, and ultimately, patients.

This confidence hinges on a collaborative ecosystem, wherein reliable infrastructure, secure AI, and validated applications function seamlessly together. That’s why we are committed to Cisco Compatible Solutions for AI. By integrating a comprehensive range of Cisco infrastructure—with compute solutions powered by Intel and other contributors—alongside a growing suite of independent software vendors, we are enabling healthcare providers to implement AI solutions that are proven to work in harmony, thereby fostering confidence, trust, and ease of use in every deployment.

The future of healthcare AI will be shaped by countless trusted decisions made daily, backed by technology specifically designed to support clinicians, healthcare organizations, and patients alike.

To discover more about how Cisco Compatible Solutions for AI integrates validated partner solutions with Cisco's infrastructure to expedite AI implementation, visit the Cisco Compatible Solutions for AI page.

Healthcare AI Needs to Build Trust Before Being Adopted

A hospital consultation room bathed in warm natural morning light. A senior doctor in a white coat sits across from an adult patient at a polished desk. Between them, a sleek touchscreen monitor displays a softly glowing abstract neural-network icon alongside patient vitals. The doctor’s face is thoughtful and cautious as they gently point to the data, while the patient meets their gaze with hopeful uncertainty. The scene conveys the delicate journey of healthcare AI earning trust.

Healthcare AI Must Earn Trust Before Adoption - Cisco Blogs

- Trust in AI for healthcare is essential and must be established at multiple levels.  
- Sickbay consolidates patient data for clinicians, enhancing decision-making and reducing alarm fatigue.  
- Collaboration with Fortanix ensures AI operates securely while protecting sensitive healthcare data.  

Healthcare has never lacked innovation.

The industry has embraced electronic health records, connected medical devices, robotic surgery, advanced imaging, and now artificial intelligence. But unlike many industries, healthcare doesn’t adopt technology simply when it’s possible. It adopts technology when it can be trusted.

That’s why the conversation around AI is changing.

The question is no longer, “Can AI effectively enable clinicians?” Increasingly, the question is, “Can we trust AI enough to make it part of clinical workflows?”

That trust isn’t built through a single model or application. It’s earned across every layer of the technology stack, from the data clinicians rely on to the infrastructure protecting sensitive information and the workflows that ultimately impact patients. That’s the thinking behind Cisco Compatible Solutions for AI. By bringing together Cisco infrastructure with an expanding ecosystem of validated software partners, we’re helping healthcare organizations deploy AI with greater confidence while building trust into every layer of the solution.

Trust Begins With Better Clinical Insight

Every second, hospitals generate an enormous volume of physiological data from monitors, ventilators, infusion pumps, and other bedside devices. The challenge is bringing that data together in a way that helps clinicians access data to inform their assessment of patient status.

That’s what makes Sickbay so compelling.

Running on Cisco Unified Edge, Sickbay brings high-frequency patient monitoring data from a wide range of bedside devices into a single view. Instead of clinicians jumping between multiple systems, they get one dashboard that helps them see the full picture, with proactive alerts that can streamline data visibility for care teams.

By bringing together continuous waveform data with clinical and electronic health record (EHR) information, AI can assist in identifying patterns for clinical review. That helps reduce alarm fatigue, supports timely access to information, and gives clinicians greater confidence in the decisions they make.

Trust Depends on Protecting Sensitive Data

As AI becomes part of clinical workflows, healthcare organizations face another challenge: protecting some of the world’s most sensitive information while enabling innovation.

Healthcare has always required strong security, but AI introduces new considerations. It’s no longer just patient records that need protection. Models, prompts, inference data, and intellectual property all become part of the attack surface.

That’s where our work with Fortanix represents an important step forward.

Together, Cisco and Fortanix are helping organizations run AI workloads inside confidential computing environments, where models and sensitive data remain protected even while they’re actively processing information. Hardware-based isolation, attestation, and zero-trust principles provide organizations with greater confidence that AI can operate on sensitive data without exposing it to unnecessary risk.

For healthcare, trust isn’t simply about keeping data encrypted. It’s about providing infrastructure designed to support the secure operation of AI where sensitive information exists.

Trust Extends to Everyday Care

Some of the most meaningful aspects of clinical care and recovery can happen outside the clinical and operational environments.

Consider something as simple and important as meal delivery. A key portion of patient management and support can be ensuring that the proper dietary regimen is followed.

Proppos uses computer vision and AI to verify that a patient’s meal matches their dietary restrictions before it leaves the kitchen.

It’s a straightforward application, but it reflects something much bigger.

Healthcare AI isn’t only about supporting complex clinical assessments. It’s also about reducing preventable errors, improving consistency, and helping clinical teams focus on clinical tasks instead of manual verification.

Confidence is the Real Foundation for Healthcare AI

These examples may solve very different problems, but they all point to the same conclusion.

Healthcare organizations won’t measure the success of AI by the number of models they deploy. They’ll measure it by the confidence those solutions create for clinicians, IT leaders, and ultimately, patients.

That confidence depends on an ecosystem where trusted infrastructure, secure AI, and validated applications work together.

That’s why we’ve been investing in Cisco Compatible Solutions for AI. By combining a full stack of Cisco infrastructure, including compute solutions powered by Intel and others, with an expanding ecosystem of independent software vendors, we’re helping healthcare organizations deploy AI solutions that have been tested to work together; building great confidence, trust, and simplicity into every deployment.

The future of healthcare AI will be defined by thousands of trusted decisions made every day, supported by technology designed to assist clinicians, healthcare organizations, and patients.

To learn more about how Cisco Compatible Solutions for AI brings together validated partner solutions with Cisco infrastructure to accelerate AI deployment, visit the Cisco Compatible Solutions for AI page.

AI in healthcare: Is a chatbot capable of responding to your medical inquiries?

More from AI in healthcare

Patients distrustful of governments and companies promoting AI as a solution for rural healthcare.

Why Healthcare AI Vendor Risk Demands Stronger Oversight

The Rise of Agentic AI in Healthcare Information Management

Integration of consumer AI into healthcare pathways