AI hallucinations infiltrate Australian parliament policy submissions

A Guardian Australia analysis reveals that dozens of policy submissions to Australian parliamentary inquiries contain fabricated research and nonexistent citations generated by artificial intelligence. The investigation found that large language models are inventing studies and falsely attributing findings to real academics across the political spectrum. In some instances, official committee reports have cited these submissions despite the majority of their sources appearing to be AI hallucinations. This infiltration compromises the integrity of the government inquiry process, which relies on accurate expert testimony and constituent feedback to inform legislation. The findings demonstrate how generative AI can systematically degrade democratic institutions by introducing plausible-sounding but entirely fictitious evidence into formal policy deliberations. Neither specific committees nor affected academics were named in the available excerpt, though the report characterizes the trend as widespread and politically agnostic.