AI-generated references raise concerns over scientific accuracy

Professor's scientific paper discovered filled with errors due to reliance on AI-generated references.

A Danish biologist, Henrik Enghoff, from the Natural History Museum of Denmark, was taken aback when he discovered his name mentioned multiple times in a scientific paper related to millipedes, referencing papers that simply didn't exist. This case has sparked concerns over the increasing use of AI, like OpenAI's ChatGPT, in scientific research.

As reported by Retraction Watch, Enghoff suspected that the authors of the paper from China and Africa might have used OpenAI's ChatGPT to find academic references, and his suspicions were proven correct.

The problematic paper was initially removed by Preprints.org, an academic preprint archive managed by the publisher MDPI, in June, following notification from Enghoff's colleague, David Richard Nash from the University of Copenhagen, who pointed out the numerous errors.

The incident has raised questions about the integrity of scientific research and how much reliance is placed on artificial intelligence tools to create links and content. This emphasizes the importance of maintaining confidence in the scientific literature, as such errors can undermine the credibility of scientific papers.