The integration of AI chatbots into the healthcare landscape is on the rise, impacting patient care both inside and outside medical facilities. A recent survey from the Pew Research Center reveals that both patients and healthcare professionals are increasingly utilizing AI for diagnosing conditions and recommending treatments, yet concerns about transparency in these interactions persist.
Conducted from June 22 to June 28, the survey involved 3,488 U.S. adults and found that around 25% of participants have employed chatbots to help diagnose their symptoms. The primary motivations for using these tools include their convenience and cost-effectiveness. Many patients turn to AI chatbots for quick access to information, clarification of a healthcare provider’s diagnosis, better comprehension of lab results, and assistance in determining if they need to seek in-person medical attention.
The responses indicate that nearly half of those surveyed (47%) found the health information provided by chatbots to be "extremely" helpful, while 48% found it "somewhat" beneficial. A mere 5% reported that the advice they received was not helpful at all.
AI chatbots function by mimicking conversational exchanges, utilizing extensive datasets from which they have been trained to formulate responses to users’ inquiries. Many users appreciate the 24/7 availability of these chatbots, as they can often access them free of charge and without the need for an account. However, there are concerns regarding their reliability, as chatbots can produce misleading information—often referred to as "hallucinations." The tendency of AI to provide plausible-sounding answers without necessarily being accurate can be especially troubling in health-related contexts.
In physician offices, the role of generative AI is expanding as well. According to Pew's findings, while many Americans have used chatbots for health-related inquiries, there is escalating anxiety concerning AI's application in clinical settings.
In the realm of physicians, the survey highlights that a significant portion of patients—72%—believe it is crucial for healthcare providers to inform them about any use of AI in their treatment process. Moreover, around 80% of respondents expect to be notified if AI influences their care or medical decisions, including diagnoses or the evaluation of lab tests and imaging results. AI technologies are often employed for triage and have demonstrated superior efficiency to human emergency room physicians, reducing the time required for X-ray assessments from an average of 11.2 days to just 2.7 days.
Despite a growing reliance on AI in healthcare, nearly half of the respondents (47%) were uncertain about whether AI had played a role in their treatment, highlighting a significant gap in patient awareness regarding the application of this technology.
While the convenience of AI chatbots is acknowledged, the survey indicates a lack of full trust among Americans. Comfort levels vary when it comes to sharing personal health information; with only 29% feeling very comfortable doing so, and 26% expressing discomfort.
Thus, even as AI becomes more prevalent in healthcare, public confidence and understanding have not evolved at the same pace as its implementation.
