TLDR: A recent study by Johns Hopkins University researchers indicates that doctors who rely on generative AI for medical decision-making are often viewed negatively by their peers, who perceive them as less competent. This ‘competence penalty’ is most pronounced when AI is used as a primary tool, suggesting a significant social barrier to AI adoption in healthcare despite its recognized benefits for accuracy.
A groundbreaking study conducted by researchers at Johns Hopkins University has shed light on a critical social barrier to the widespread adoption of artificial intelligence in healthcare: peer perception. The study, published in the August issue of npj Digital Medicine, reveals that physicians who integrate generative AI into their medical decision-making processes are frequently viewed with skepticism and deemed less competent by their colleagues.
The research involved a randomized experiment with 276 practicing clinicians, including 178 physicians, who evaluated different clinical scenarios. These scenarios presented a physician using no AI, one using AI as a primary decision-making tool, and another using AI for verification purposes. The findings consistently demonstrated a ‘competence penalty’ for AI-dependent physicians.
Specifically, physicians who relied on AI for primary decision-making received significantly lower ratings for clinical skills, competence, and overall healthcare experience. On a scale of 1 to 7, those using AI as a primary tool scored a mean of 3.79 for clinical skill, starkly contrasting with the 5.93 mean score for physicians who did not use AI at all. When AI was used for verification, the perception improved slightly, with a mean score of 4.99, but still fell short of the non-AI group. Similarly, in assessments of ‘perceived professionalism,’ the AI-dependent group scored 3.71, compared to 4.94 for AI-verification users and 5.99 for non-AI users.
Tinglong Dai, a professor of business at the Johns Hopkins Carey Business School and co-corresponding author of the study, expressed surprise at the findings. “What surprised us is that doctors who use it in making medical decisions can be perceived by their peers as less capable. That kind of stigma, not the technology itself, may be an obstacle to better care,” Dai stated.
Haiyang Yang, the first author of the study and academic program director of the Masters of Science in Management program at Carey, emphasized the enduring role of human psychology. “In the age of AI, human psychology remains the ultimate variable. The way people perceive AI use can matter just as much as, or even more than, the performance of the AI itself,” Yang commented.
The implications of this study are substantial. Despite the immense promise of generative AI in advancing healthcare and the increasing pressure on clinicians to adopt these technologies – with McKinsey reporting that roughly 85% of healthcare leaders were already using or exploring generative AI by the end of 2024 – this social stigma could significantly hinder its integration. Risa Wolf, a senior author from Johns Hopkins University, noted, “As AI tools become more commonly used in healthcare and in medicine, I think this really just demonstrates that there are going to be challenges, some barriers to adoption and increasing use. But it also highlights that there need to be thoughtful approaches to its implementation.”
Also Read:
- Clinical AI Poised for Explosive Growth to $188 Billion by 2030, Navigating Key Hurdles
- Harvard Study Uncovers Practical AI Integration Among Students and Professionals
Ironically, while peer perception suffers, the study also found that clinicians generally recognize the benefits of AI. Participants rated generative AI as useful for ensuring the accuracy of a physician’s clinical assessment, with an average score of 4.30 (out of 7) overall, and 4.96 for AI tools customized to their institution. This suggests a disconnect between the perceived utility of AI and the social implications of its use, highlighting a complex challenge for the future of AI in medicine.
