The Report the Government Used to Ban Kids From Social Media Has Fake References in It

The report that the Albanese government relied on to justify banning under-16s from social media contains references to academic studies that simply do not exist, and at least some look like they were made up by AI, according to a Guardian Australia analysis published on August 17, 2026 (Guardian).

The report was produced by a UK-based company called the Age Check Certification Scheme, or ACCS, at a cost of $3.48 million. Communications minister Anika Wells said it showed there were "many effective options" for checking people's ages online. It helped pave the way for the ban to start on December 9, 2025, making Australia the first country in the world to enforce such a restriction (Reuters). Reuters reported that 86% of Australians aged eight to 15 used social media.

The problems are extensive. Guardian Australia found six references in one section of the ACCS report containing errors. A submission to an ongoing Senate inquiry flagged at least two citation errors and said the references "appear to be AI hallucinated, rather than being based on real sources" (Guardian).

When AI programs like ChatGPT are asked to write about a topic, they sometimes invent sources that look real but are completely made up — fake author names, fake journal titles, fake studies. That is what "AI hallucination" means. The errors in the ACCS report take several forms. Some reference links point to academic papers that do not exist. Other links go to real papers, but not the ones the report claims to be citing. Author names, journal titles, and publication years are mixed together in ways that do not match any real research. In some cases, a link leads to a genuine paper, but the paper does not say what the ACCS report claims it says.

One study cited in the report was apparently accessed in March 2025, but the lead author told the Guardian the paper was not publicly available until it was published in June 2025. The same author said "Jamil", whom ACCS listed as the paper's lead author, had never been part of the paper, and that the ACCS report had described the work incorrectly. Another reference, "Weber et al. 2011", could not be found in the journal cited. An ACCS spokesperson said the reference came from a Monash University study, but the Monash paper ACCS supplied contained no reference attributed to Weber.

Then it got worse for ACCS. The spokesperson initially denied any use of AI in producing the report. That became impossible to maintain when the Guardian found four links in the report containing hidden metadata identifying ChatGPT as their source. Only then did the spokesperson admit AI had been used, though the admission was narrow: AI was used to rewrite paragraphs more concisely. The spokesperson denied the citation errors were caused by AI.

ACCS also argued that having ChatGPT metadata in the links counted as disclosing AI use, and since the links worked, there was no problem. That is a generous reading of transparency. A person reading the report would have to dig into the underlying code of a hyperlink to find that metadata, which is not the same as openly stating in the report that AI was used.

The attempted fixes made things worse. The replacement references ACCS supplied to correct the problematic citations also contained errors, with wrong years, authors, and journal titles. For a report meant to prove that age-checking technology is reliable, not being able to get its own references right is a pointed irony.

The broader context here matters. The trial, assessed by a company called KJR (KJR), tested various ways of checking or estimating someone's age online. The final report, published by the Department of Infrastructure in August 2025 (Infrastructure.gov.au), found that some errors in age checking are inevitable (Guardian, August 2025). The government pressed ahead anyway, announcing it would introduce age restrictions after the report came out (Biometric Update).

What is now in question is not just one flawed document but the whole chain of evidence behind a world-first law. The ACCS report was the technical basis ministers pointed to when arguing the ban could work. If that basis contains fabricated citations and AI use that was not disclosed, the scrutiny now coming through the Senate inquiry is both predictable and overdue.

The ACCS trial has formally ended and is no longer accepting new entries (ageassurance.com.au). The ban it supports has been running since December. Parliament cannot undo a law that is already in force, but the Senate inquiry now has a straightforward question to answer: whether the government passed a world-first restriction based on a report whose own authors could not keep their references straight.