During a recent Senate inquiry, questions were raised regarding the integrity of a report used to support the Australian government’s proposed ban on social media for teenagers. The inquiry focused on the discovery that several academic citations referenced within the report do not appear to exist.

When questioned about the source of these errors, the report's authors acknowledged that ChatGPT was utilized to rewrite and shorten sections of the text. However, the authors explicitly denied that the citation errors were the result of 'AI hallucinations'—a term used to describe instances where artificial intelligence generates false information.

This admission follows a period of shifting explanations regarding the use of technology in the report's production. A spokesperson for the authors initially denied the use of AI when first approached by media outlets earlier this month. The spokesperson later revised this statement, confirming that AI tools were indeed employed to improve the conciseness of the writing. The Senate inquiry is continuing to investigate the extent of AI involvement in the document and whether the reliance on such tools has compromised the validity of the evidence presented to lawmakers.