Is AI Flooding Medical Journals With 'Meaningless' Research?
(MedPage Today) -- As artificial intelligence (AI) grows more sophisticated, its use in medical research is becoming both more prominent and harder to detect, leading to what some describe as a flood of mediocre science being published in journals...
In recent months, artificial intelligence (AI) has become increasingly prevalent in medical research, raising concerns about the flood of low-quality studies being published in journals. While earlier this year, scientists were worried about AI-generated fabricated references, it's now relatively straightforward to use AI to produce research without noticeable errors, according to experts.
Detecting the presence of AI in research is becoming more challenging, as Elisabeth Bik, PhD, a science integrity consultant, states that at this point, it's impossible to distinguish between fake and real studies. Ivan Oransky, MD, co-founder of Retraction Watch, explained that creating studies with findings of minimal consequence is now relatively easy using AI tools like ChatGPT or Claude. Systematic reviews are particularly susceptible to manipulation through AI.
The use of public datasets, such as the CDC's WONDER or the National Health and Nutrition Examination Survey, has surged since the introduction of generative AI. Most biomedical publications now exhibit signs of AI-assisted writing, according to a study cited by research integrity expert Matt Spick, PhD, from the University of Surrey in England.
Some publishers have responded to the influx of AI-generated studies by implementing new strategies. Frontiers, an open-access publisher, required new experimental validation or additional data from authors' institutions for specific types of manuscripts. Over 12,000 submissions were rejected by Frontiers' research integrity team in the past year due to simple queries of public datasets, including a significant number of NHANES-based submissions.
Other journals are also reviewing their policies to ensure the flood of mediocre studies doesn't overwhelm the peer review system. Springer Nature and NEJM Group are piloting AI detection tools and improving disclosure efforts. However, these tools may struggle to keep up with AI's constant improvements.
The rise of AI in medical research also raises concerns about outright fraud. Generative AI has made it easier to create fraudulent studies, which authors are better equipped to respond to. As long as the competitive pressure to publish persists in academic medicine, journals are likely to continue receiving an overwhelming amount of low-quality AI-created work, experts suggest.
Written by urgent.news from MedPage Today's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
