EU begins enforcing AI Act as new transparency rules take effect
As of last Sunday, the European Commission’s AI Office and EU Member States’ national authorities have begun enforcing the AI Act, requiring AI systems to tell when users are interacting with AI and when content has been generated or altered by AI.
Generative AI is getting better and better at creating and altering texts, photos, videos, and voices. In fact, sometimes, the content that’s been produced by AI is virtually indistinguishable from its real-life counterparts.
As a result, it’s becoming harder and harder to tell when we’re dealing with authentic content and real human beings, or when AI is being used.
This opens the door to all sorts of risks, including identity fraud, online scams, misinformation, and other forms of deception.
That’s why the EU launched the AI Act in August 2024, the world's first comprehensive law on artificial intelligence. The legislation is implemented in stages. On Sunday August 2nd, 2026, the AI Act’s transparency rules went into effect.
Under these new rules, chatbots and other interactive systems will have to inform users whether they are dealing with an AI system or human being. Furthermore, deepfakes will have to be labelled. Any other content that has been AI-generated or altered by AI has to carry a digital watermark that can’t be removed or cropped.
Lastly, if a company uses AI to analyze emotions or categorize people based on biometric characteristics, people have to be informed at the first moment of interaction.
“The measures are intended to reduce deception and manipulation and help people make informed choices. They also give businesses clearer obligations and a practical way to show compliance,” the European Commission says in a press release.
Additionally, it helps to ensure that AI is developed, deployed, and used safely, giving both people and businesses greater confidence in the capabilities and opportunities of the technology.
“With the AI Act, we established a clear, risk-based and durable framework for trustworthy AI, one that gives innovators legal certainty while protecting the public interest. As enforcement begins, we are taking an important step towards AI that people and businesses can understand and trust, and whose benefits are shared widely across our society,” Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy, says in a statement.
The application of the rules on high-risk AI systems integrated into regulated products, like smart cameras equipped with facial recognition or software designed to assess applicants, has been postponed to August 2nd, 2028.
The rules on AI systems that generate non-consensual sexually explicit content and child sexual abuse material will apply from December 2nd, 2026.