Retraction Watch's weekly roundup touches on several converging threads: calls to sanction authors who use AI-hallucinated references, a striking case where an AI tool flagged plagiarism in 95% of PhD theses at an Indian university, and broader questions about accountability in an era of AI-assisted research fraud. The sanctions angle is particularly significant — it represents the move from detection to consequences, a policy frontier that scholarly publishing has not yet resolved.

Source: Retraction Watch