Key Insights
- Researchers are increasingly using artificial intelligence models to plumb scientific literature.
- The models often do not distinguish between solid research and research that is dubious or has been retracted.
- Scientists have proposed several ways to confront the problem, but they will require the cooperation of academic publishers.
Artificial intelligence agents are increasingly being used to interpret and even conduct science. But are they able to accurately distinguish science that is solid from science that is subpar?
Valentin Rodionov wanted to find out. To do so, he chose 42 studies that don’t meet the parameters of legitimate research. Some have been retracted from literature; others are considered to be problematic, fraudulent, or pseudoscientific but haven’t formally been retracted or corrected.
Valentin Rodionov, an organic chemist at Case Western Reserve University, was surprised at how uncritical artificial intelligence models are about dubious research. Credit:
Courtesy of Valentin Rodionov
