Australia age-verification report hit by AI error claims

Australia’s planned social media ban for users under 16 is facing fresh scrutiny after questions were raised about the reliability of a government-commissioned report on age-verification technology. Critics say parts of the document include inaccuracies that resemble problems linked to automated text generation rather than conventional research errors.

The report cost $3.48 million and is described as a key technical plank supporting the policy case for restricting under-16 access. The current dispute centres on whether the document’s underlying references and technical identifiers are sound enough to support legislation that is already under active examination.

Disputed citations and identifiers raise integrity questions Those Disputed citations and identifiers raise integrity questions Those challenging the report say it contains multiple Those challenging the report say it contains multiple academic citations that cannot be located in academic databases. They also point to digital identifiers that do not correspond to any matching records, suggesting the referencing and sourcing may not be dependable. The disputed elements have fuelled an allegation that the errors look like common “hallucination”-style problems associated with automated text tools. The concern, as framed by critics, is not simply a few broken references but the possibility that the evidentiary foundation is weaker than it appears. Contractor shifts position after metadata is reported The contractor initially rejected suggestions that artificial intelligence was used to produce the report. After subsequent analysis identified metadata described as generated by ChatGPT within the document, the integrity debate intensified.

Following the metadata findings being reported, the firm said AI had been used only for editorial work. It maintained that the research and substantive content were completed by humans, and that all research was human-verified.

The contractor said the problematic citations and identifier mismatches were the result of broken links, not fabricated references. Critics, however, have continued to question that explanation given the number of disputed items and the presence of AI-linked metadata.

Officials acknowledge concerns as Senate inquiry reviews the

Officials acknowledge concerns as Senate inquiry reviews the ban Officials have acknowledged the concerns and said they are engaging with the contractor to address discrepancies. The department responsible has not independently confirmed the report’s accuracy.

Officials also pointed to the practical difficulty of auditing technical claims after the fact, describing retrospective verification as challenging. That position has added political sensitivity because the report is part of the technical basis for the under-16 restriction.

The controversy is unfolding while a Senate inquiry examines the proposed social media ban. The inquiry is considering how the disputed parts of the report affect the legislation’s evidentiary grounding, although officials have not said the legislation will change.

The episode also reflects earlier government scrutiny over how external consultants use AI in official work. For lawmakers and the public, an open question remains: whether the report’s technical conclusions can still be treated as dependable given the issues around references and identifiers.