Belgian ad watchdog sanctions AI-generated poster found misleading: a wake-up call for businesses

On 13 October 2026, Belgium's Advertising Ethics Jury (JEP) ordered the withdrawal of an in-store poster advertising a product as "zero sugar" when it actually contained cane sugar. The retailer admitted the error came from an AI-generated layout. The ruling lands weeks after the EU AI Act's transparency obligations for AI-generated advertising content took effect in Belgium on 2 August 2026. For Belgian businesses, the message is clear: producing content faster does not remove the need to check it.

1. What the JEP found

A customer flagged an in-store poster for an energy drink claiming "zero sugar, zero calories, source of caffeine". Yet the second ingredient listed was cane sugar. The JEP ruled the ad misleading under articles 4 and 5 of the Belgian Advertising Code, which forbid misleading the average consumer about a product's characteristics.

When questioned, the retailer admitted the mistake: the poster had been generated by artificial intelligence, and no one had cross-checked the generated text against the product's actual composition. The retailer confirmed the poster's removal following the decision.

2. A legal framework tightening at the same time

Since 2 August 2026, the EU AI Act has required advertisers, agencies and other communication professionals to clearly identify any advertising content generated or manipulated by AI that could be mistaken for authentic content, according to the Belgian Advertising Council. The disclosure must be noticeable from first exposure, without the public needing to search for it.

This transparency obligation does not replace existing rules on fairness and accuracy in advertising. The "zero sugar" poster case shows both problems can stack up: AI-generated content that was not flagged as such, and that was also factually wrong.

3. Why this matters to any Belgian business that communicates

The JEP handles dozens of cases every month, and most ads reviewed raise no issue. But this case illustrates a risk specific to AI-generated content: the tool produces a plausible, polished text that can drift from the product's actual facts without any warning sign appearing.

For an SME using AI to produce campaign visuals or copy, vigilance must cover both substance (is the displayed information accurate) and form (is AI-generated content flagged when regulation requires it).

The AI lens, humans first

An image or copy generator executes a brief, it does not check a product's composition or the accuracy of a claim. It can produce a visually consistent poster while repeating false or outdated information, simply because nothing in its brief contradicted it. The decision to publish, and the proofreading that checks every claim against the product's reality, remain a human task. AI executes. Expertise decides, and keeps watch.