Swedish AI Expert on the Chopping Block After ChatGPT Fakes 40+ References in Her Own Book
A top Swedish AI researcher is now under official investigation for scientific misconduct after her new book cited dozens of references that were completely fabricated by an AI tool.
The researcher, a prominent voice in responsible AI, used a generative program to compile her source list, but the software “hallucinated” several citations that do not exist. Sweden’s National Board for Assessment of Research Misconduct (NPOF) is now reviewing the case to determine if the incident was deliberate deceit or simple negligence [1]. Jonas Jeppson of NPOF confirmed the agency is seeing a spike in such cases, directly linking the rise of automated writing tools to a new wave of academic integrity failures [1]. A formal finding of misconduct could damage the researcher’s credibility and career, though no final decision has been made public [1]. The case underscores a growing problem across academia: as AI tools become standard for drafting and sourcing, the line between assistance and error is blurring—with potentially career-ending consequences.