The European Commission has begun actively enforcing a sweeping new set of transparency rules under the EU’s Artificial Intelligence Act, marking one of the bloc’s most consequential steps yet in regulating how AI systems interact with the public. The rules, which apply to companies both inside and outside the EU that serve European users, carry financial penalties steep enough to force even the largest technology firms to take notice.
What the New Rules Require
Under Article 50 of the AI Act, providers of AI systems that interact directly with people, including chatbots, voice assistants and AI agents, must now clearly disclose that users are engaging with artificial intelligence rather than a human, unless that fact is already obvious from the context. The obligation extends to content generation as well: any AI-generated or manipulated image, audio or video, commonly known as a deepfake, must be labelled, while text, audio and visual content produced by AI systems must carry machine-readable marks that allow it to be identified more easily by platforms and users alike.
The rules distinguish between two categories of responsibility. Providers, meaning the companies that build and place AI systems on the market, bear the primary obligation to ensure their tools disclose their artificial nature. Deployers, the businesses and individuals who put those tools to use, carry additional responsibilities around labelling AI-generated content shared with the public, particularly in contexts involving public interest information.
Notably, the obligations are not retroactive. Content generated or published before the rules took effect does not need to be relabeled after the fact, though a limited transitional period applies specifically to the marking and detection requirements for generative AI systems already on the market, with full compliance required by early December.
Steep Penalties for Non-Compliance
The financial stakes attached to non-compliance are substantial. Violations of the AI Act’s transparency obligations can trigger fines of up to €15 million or 3 percent of a company’s global annual turnover, whichever figure is higher, while breaches of the Act’s more serious banned-practice provisions can reach as high as €35 million or 7 percent of worldwide turnover. Given the scale of revenue generated by major AI providers, the penalties are designed to function as a genuine deterrent rather than a routine cost of doing business.
To help companies navigate the new requirements, the European Commission’s AI Office has published a voluntary Code of Practice on Transparency of AI-Generated Content, including a standardised set of icons that providers can use to label their outputs. Several major AI companies have already signed on to the code, which offers participating firms a degree of presumption of conformity along with a more favourable regulatory posture should disputes arise.
A Parallel Push to Expand AI Infrastructure
The enforcement drive has landed in the same week as a separate, seemingly contradictory move by Brussels: a call for tenders to establish up to seven AI “gigafactories” across Europe, aimed at dramatically expanding the continent’s AI computing capacity. The initiative pairs roughly €10 billion in public funding from the EU and member states with an expected €20 billion or more in private investment, bringing the total potential investment above €30 billion.
The gigafactories are intended to give European researchers, companies and public institutions access to the kind of large-scale infrastructure needed to train and run advanced frontier AI models, an area where Europe has historically lagged behind the United States and China. The tender window remains open until mid-November, with award decisions expected in early 2027.
Regulating and Investing at the Same Time
Taken together, the two initiatives illustrate the balancing act Brussels is attempting: tightening the rules governing how AI systems disclose themselves to the public, while simultaneously investing heavily to ensure European companies are not left behind in the broader global AI race. For companies operating in the EU market, the message from Brussels this week has been unambiguous: transparency obligations are no longer aspirational guidance but an active compliance requirement, backed by fines large enough to reshape corporate risk calculations, even as the bloc works in parallel to build the infrastructure needed to keep Europe competitive in the AI economy going forward. Next Article
