Report on Australia teen social media ban contains AI citation errors, Senate hears

Guardian analysts have identified a flaw in a recent report on Australia’s under‑16 social‑media ban, noting that one section contains links to academic articles that cannot be located. The report, which evaluated a $3.48 million age‑assurance technology trial conducted last year by the UK‑based Age Check Certification Scheme (ACCS), was intended to assess tools that could verify users’ ages for platforms subject to the new legislation. While the authors acknowledge that ChatGPT was employed during the editing process, they reject claims that the erroneous citations stem from AI‑generated hallucinations, insisting the references were not fabricated by the system.

The ACCS trial examined a range of verification methods, from document‑based checks to biometric solutions, to determine their suitability for enforcing the ban. Despite the controversy over the report’s citation accuracy, officials say the findings will still inform policy decisions on how social‑media companies can comply with Australia’s age‑restriction framework. The dispute highlights ongoing concerns about the reliability of AI‑assisted research tools in high‑stakes regulatory contexts.