The UK’s Advertising Standards Authority has upheld a complaint about a paid Facebook ad for AI Scorecast, an app promoted for predicting football results. The ruling, published on 14 October 2026, concerns HRY Games, trading as AI Scorecast.
The claim behind the ruling
The ad, seen in April, claimed 97% accuracy. The ASA found that it implied users were very likely to profit from following the recommendations while omitting the possibility of losing money. It judged the ad misleading and irresponsible.
HRY Games did not respond to the investigation. The ASA said the ad could not run again in that form. The ruling does not establish a verified 97% performance result or ban the app itself.
Assessing the message as well as the model
For iGaming marketing teams, the practical issue is how a technical figure changes meaning when placed beside a betting proposition. A reader may understand a model statistic as reassurance about their own money, even when the advertiser intended to describe a narrower prediction task.
An editorial review could start with the definition behind the number: what counts as a correct prediction, which matches and period were tested, and whether unsuccessful outputs remain in the sample. Reviewers could also check whether the test reflects the conditions in which customers would use the product.
The next step is to read the finished creative without the supporting presentation. Headlines, testimonials and calls to action together create the message a player actually receives. A technically precise metric can still be used in a way that invites a much broader conclusion. Keeping dated evidence alongside each creative version gives reviewers a concrete basis for revisiting a campaign when the product or its claims change.
Our coverage of KVA’s study of Dutch casino licence errors explored another claim-verification problem in player-facing AI. For promotional materials, the comparable task is to keep the evidence, the model’s stated purpose and the advertised promise aligned.