- Published on
Report supporting Australia’s teen social media ban appears to contain AI hallucinations, Senate hears
- Authors
- Name
- DeepHorus
The Senate has heard that a report supporting Australia's under-16 social media ban contains apparent AI hallucinations, as the authors conceded ChatGPT was used in editing but denied the citation errors stemmed from AI. The admission came during an ongoing inquiry into the ban, which took effect in December last year. The report's credibility is now under renewed scrutiny.
The $3.48 million age assurance technology trial was run by the UK-based Age Check Certification Scheme (ACCS) and tested various technologies for social media platforms. Communications minister Anika Wells heralded the report as showing “many effective options” for age checking, paving the way for the ban. Most of the report focused on existing technologies, but one chapter covered “emerging technologies” and cited numerous academic papers.
Guardian Australia's analysis identified six references in that chapter containing errors. These included Digital Object Identifiers (DOIs) linking to non-existent papers, DOIs pointing to incorrect papers, and author, journal, and publication year combinations that matched no known references. Some DOI links led to papers that did not say what the report suggested.
One cited article was apparently accessed in March 2025, but a lead author confirmed it was not publicly accessible until June 2025. The author also said that “Jamil,” named as the paper's lead author, had never been on the paper. Another reference, “Weber et al. 2011,” could not be found in the cited journal, nor in a Monash University study provided as support.
ACCS initially denied using AI, but later conceded ChatGPT was used to rewrite paragraphs more succinctly. A spokesperson insisted AI was not used for originating research or generating the report, and that all links were human-verified. They apologized for any inadvertent errors but maintained the links' ChatGPT metadata disclosed the tool's use.
The federal government has previously taken a dim view of contractors using AI in reports, with Deloitte refunding part of a $440,000 contract last year. The revelations raise questions about the reliability of evidence used to justify the ban. The inquiry is likely to probe how such errors slipped through.
As the Senate inquiry continues, the flawed citations may prompt further review of the report's conclusions. ACCS has provided new references, but those too contained mismatches. The episode underscores the need for rigorous verification when AI tools are used in policy-shaping documents.
