Science, but not our tipping work is being blighted by AI.

Day 3 of York's Dante Meeting

Research conducted at a British university has uncovered a worrying trend: the proliferation of low-quality, AI-generated research papers that could undermine genuine scientific understanding. The study, published in the open-access journal PLOS Biology, details an “explosion of formulaic research articles” that demonstrate inappropriate methodologies and report false findings. By analysing papers that utilized data from the US National Health and Nutrition Examination Survey (NHANES), the University of Surrey team observed that many publications after 2021 employed “a superficial and oversimplified approach to analysis.” This often involved examining single variables in isolation, neglecting the complex interplay of factors influencing health, and sometimes involved the unjustified selection of limited data. According to Dr. Matt Spick, a lecturer in health and biomedical data analytics at Surrey University and a report author, this influx of scientifically unsound papers represents “science fiction” that deceptively uses national health datasets.

I know what you’re thinking. Why does the phrase “…a superficial and oversimplified approach to analysis…” resonate? I don’t think anyone can suggest that I don’t understand the “complex interplay of factors” when it comes to in-depth race analysis. It’s just that sometimes, life gets in the way, and I have to resort to using a pre-chip and e-anything technology to complete my complex analysis.

For those of you of a scientific bent, this is a full and frank disclosure of one of my many analytical and selection methods, which has delivered results entirely in line with the mathematical odds.