The EMIF-funded Unreliable Science Project investigated how retracted papers continue to circulate in public discourse, including in the media and on social media, even after retraction, and how they may contribute to misinformation and disinformation. The project also examined how large language models (LLMs) deal with retracted papers and whether they continue to treat them as credible scientific evidence despite their retracted status. In addition, it addressed the dissemination of problematic scientific sources such as retracted papers and pseudoscience from a legal perspective, including questions of freedom of speech and media law.

Big thanks to Thomas Stollenwerk and Michael Anranter for presenting findings to our research group. Little surprise: LLMs were not found to be the ones solving the problem of the long afterlife of retracted publications.