A recent Medscape report revealed that 24% of respondents use AI tools to analyze medical literature and data. While these AI tools can help organize data and generate content, they still lack reliable references, which may result in misleading information.
According to Corey S. Maas, MD, facial plastic surgeon and founder of The Maas Clinic in San Francisco, “The real problem right now is to assume that the information delivered by ChatGPT is correct, given that the information is so well-written, but that’s not always the case.” Dr. Maas emphasized the importance of verifying AI-generated content to avoid inaccuracies that could compromise the integrity of research.
The report also highlights the role of specificity in enhancing the performance of AI models. Precise queries can improve AI output, but the tools remain susceptible to “hallucinations,” where false or fabricated information is generated. Warren D’Souza, PhD, co-director of the University of Maryland Institute for Health Computing in Baltimore, warned, “There is a significant amount of risk involved with using AI for medical research.” These AI tools are trained on a vast array of information, and D’Souza explained, “It’s hard for them to discern what’s good information and what’s bad information since it’s ingesting all of it. That’s why there should be a level of caution providers should exert when accessing such tools for research.”
To maintain research quality, trusted platforms such as PubMed remain essential. Unlike AI tools, these established resources provide rigorously vetted, peer-reviewed content that is crucial for supporting hypotheses and ensuring accurate background research.
While AI can help summarize and filter large amounts of data, its reliability is still evolving. Researchers are encouraged to rely on proven methods and trusted resources to ensure accurate and credible results.
