Gambling education pitch flagged for AI "slop" - what educators should do before buying programs
A University of Sydney-based group, the OurFutures Institute, sent politicians an "evidence review" to support a $20m bid for a school gambling prevention program targeting 15- to 20-year-olds. After scrutiny, the document showed broken links, references to studies that don't exist, and claims that clashed with the sources cited.
Independent senator David Pocock said he was "deeply concerned" and that the review "appears to just be slop written by AI," citing "AI hallucinations" and false statements. He argued the review exaggerated the case for school programs and urged the government to act on gambling ad bans that affect children.
What happened
The review reportedly contained at least 21 citation issues: dead links, non-existent papers, and mismatched sources. One headline claim - that every $1 in school-based prevention returns $8-10 in avoided health, welfare, and justice costs - does not appear in the cited Productivity Commission report. That report in fact raised reservations about school-based gambling education.
OurFutures' chief executive said the core policy rationale remains sound, attributing citation errors to a reference "editing tool" that reordered sources and created mismatches. The institute apologized, promised corrected versions, and said it would verify references line-by-line. It also stated that the proposed program will be independent of gambling-industry funding or influence.
The budget submission named leaders with program experience, including Prof Sally Gainsbury from the University of Sydney's Brain and Mind Centre. The submission did not disclose that Gainsbury receives direct and indirect funding from industry groups including Entain Australia, Sportsbet, Star Entertainment and the European Lotteries Association. Separately, researcher Samantha Thomas said two papers listed alongside her name "had not [been] written and do not seem to exist." Public health researchers Thomas and Hannah Pitt said young people consistently call for bans or strong limits on gambling ads.
Why this matters for education leaders
Education budgets are tight, and vendors know that "evidence-based" sells. AI-generated documents with confident tone and broken sourcing can slip through if no one checks the footnotes. Add potential conflicts of interest and glossy ROI claims, and it's easy to make a costly decision that doesn't help students - or that backfires.
Quick checklist before you back a gambling education program
- Open every citation. Do the links work? Does the quoted line appear in the source? Watch for DOIs that don't resolve and journals that don't exist.
- Cross-check big claims against authoritative reviews. For context, see the Australian Productivity Commission's gambling inquiry overview: pc.gov.au/inquiries/completed/gambling-2010.
- Ask for study quality. What designs were used (e.g., randomized, controlled)? Are outcomes objective or self-reported? How long was follow-up?
- Request full reference lists and access to PDFs where possible. If a paper can't be produced, treat the claim as unverified.
- Demand conflict-of-interest disclosures for all leaders, advisors, and funders. Get it in writing and publish it for your community.
- Be skeptical of ROI numbers. Ask for the full model, assumptions, and sensitivity analysis. If they can't show the math, don't cite the metric.
- Plan for independent evaluation. Pre-register outcomes, use a third party, and commit to publishing results - even if they're neutral.
- Listen to students and families. Many report harm linked to pervasive advertising; policy levers (like ad restrictions) may have more impact than classroom lessons alone.
Fast ways to spot AI-generated "slop" in proposals
- Confident tone with sweeping, unsourced claims; citations that look polished but don't resolve; copy-paste quotes that no one can find.
- Inconsistent formatting and reference styles in one document; repeated phrasing across sections; statistics with suspiciously round numbers.
- Fix: pick three bold claims and verify them line-by-line. If two fail, pause procurement until a corrected, verifiable brief arrives.
Action steps for schools and systems
- Pause any purchasing decision tied to the disputed review until corrected, independently verified materials are delivered.
- Set a district standard: no "evidence-based" claims without accessible sources, COI statements, and an evaluation plan.
- Engage independent researchers to review proposals. A 60-90 minute panel can save a year of sunk costs.
- Support students now: integrate media literacy about gambling ads, involve parents, and map referral pathways for help.
Support and resources
If you or someone you know needs help: Gambling Help Online 1800 858 858. National Debt Helpline 1800 007 007.
Level up your team's AI literacy
If your staff reviews vendor proposals or writes policy briefs, baseline AI skills matter. For practical courses by job role, see Complete AI Training.
Your membership also unlocks:
