AI labels to be compulsory on authentic-looking content…

Starting 2 August 2026, the European Union is implementing new transparency requirements under the landmark AI Act, mandating that artificially generated content designed to look authentic must be clearly labeled. These regulations require synthetic text, images, video, and audio to feature visible markers and digital watermarks to indicate their artificial origin, with specific provisions for text involving public interest that lacks human editorial oversight.

What happened

The regulation follows a surge in deceptive AI content targeting high-profile political figures. Notable examples of these deepfakes include audio recordings purporting to show Slovakian opposition leader Michal Šimečka conspiring with a journalist to rig elections, fabricated videos of French President Emmanuel Macron dancing in a 1980s nightclub, and fake footage of UK Labour leader Keir Starmer allegedly abusing party staff. These incidents, intended to malign targets and mislead the public, prompted regulators to prioritize transparency as a means of protecting democratic processes and the authenticity of online facts.

The rules apply to new AI systems entering the EU market, while existing systems have a four-month grace period for compliance. While the mandate exempts personal content, satire, and fictional works, non-compliance can result in significant financial penalties, including fines of up to €15 million or 3% of a company’s worldwide global turnover. The industry has expressed concerns regarding the breadth of these requirements, with the Computer and Communications Industry Association arguing that broad labeling could lead to user fatigue similar to cookie banners. Conversely, proponents like Green MEP Sergey Lagodinsky maintain that such transparency is essential for preserving public trust, suggesting that the industry will adapt to the requirements as they are implemented.