Crypto news

The key document that Australian authorities used as justification for banning social media for teenagers turned out to be riddled with references to non-existent scientific papers. This is an alarming signal for the entire industry: when regulatory decisions rely on data generated by neural networks, it is not only the reputation of a specific report that comes under fire, but also trust in the very process of lawmaking.

The age verification technology trials were conducted by the British organization Age Check Certification Scheme (ACCS). A total of 3.48 million Australian dollars (~$2.5 million) was spent on these efforts. Specialists tested more than 60 solutions from 48 vendors and, in August 2025, handed the government a final document of about a thousand pages across ten volumes. At the time, Communications Minister Michelle Rowland stated that the data confirmed the existence of "a multitude of effective tools" for age verification. The ban itself for users under 16 came into force on December 10, 2025.

Six references with errors

The bulk of the report is devoted to practical trials of existing solutions. Problems surfaced in the chapter on promising developments, where the authors turn to scientific literature. The first to point out the inconsistencies was the author of a submission to the Senate committee reviewing amendments to tighten the ban. The document claims that at least two references look like typical AI "hallucinations."

My review uncovered six violations in the same chapter, which fall into four categories:

• fictitious DOIs;
• DOIs redirecting to unrelated works;
• combinations of "author-journal-year" not confirmed by any publication;
• correct references to articles, but without the claims the report attributes to them.

For example, one article was dated by ACCS as March 2025, but it was actually only published in July. In another case, a reference to Weber et al. 2011 led nowhere—it was not found in the specified journal, nor in a study by Monash University.

ChatGPT in the metadata

An ACCS representative categorically denied the use of AI in preparing the document. However, after four references with metadata directly pointing to ChatGPT were found in some parts of the report, the position had to be softened. ACCS admitted that artificial intelligence rewrote certain fragments for brevity, but insisted that the research itself was created without the involvement of OpenAI's chatbot. Organization representatives stated that the ChatGPT metadata itself constitutes sufficient notice of AI usage.

However, a review of the corrected list of sources revealed new discrepancies: years, authors, and journal names in it diverged from the original citations in the report.

"A handful of errors" across a thousand pages

The Department of Infrastructure and Transport said it is examining the complaints. At Senate hearings, officials relayed ACCS's version: according to them, the references were working at the time of publication and "broke" later. First Assistant Secretary Sarah Vandenbroek clarified that the department had not independently verified this version and considers such verification "labor-intensive."

Professor Christian Downie of the Australian National University warned that false references lead to erroneous decisions and erode trust in institutions. Independent Senator Fatima Payman demanded an official apology and a refund from the contractor. She reminded that a similar incident with Deloitte a year earlier should have taught authorities to be more careful with sources. In October 2025, the consulting firm had already returned part of the payments under a contract worth 440,000 Australian dollars, admitting to the use of AI.

My expert assessment: This situation is just the tip of the iceberg. When government bodies make decisions based on reports where AI is used without proper oversight, we risk getting an entire layer of regulatory acts built on fictitious data. The industry needs strict source verification standards, otherwise trust in digital initiatives will be undermined once and for all.