As of today the EU's rules on AI models become enforceable, cementing the European Commission’s role as the world’s most prominent regulator of this disruptive technology. The AI Act, passed in 2024, introduces the first comprehensive law to regulate artificial intelligence, and its most significant provisions—especially those covering large language models—take effect this August.

The rulebook applies to all models that lack a specific purpose but can be adapted to many use cases, requiring transparency on how a model was built, disclosure of any copyright‑protected content used for training, and enough information for downstream users to understand the model’s capabilities. Companies developing the most powerful frontier models must identify and mitigate risks to society at large. Last year the Commission endorsed a voluntary code of practice drafted by world‑leading experts, including Yoshua Bengio, and most leading Western AI labs—except Meta—signed it. OpenAI’s Vice President Tom Duff Gordon said, “We have collaborated closely with the European Commission and the wider ecosystem on implementing the AI Act, including its Codes of Practice, and will continue working together to help Europe realise the benefits of the Intelligence Age.”

The Commission set up the European AI Office to enforce the Act, a huge task given the limited resources and high demand for AI talent. It will rely on external expertise, panels of scientists and specialised AI safety firms, while keeping pace with fast‑moving technological developments. Officials also face scrutiny from Washington, with the Trump administration treating EU digital rules as an attack on US commercial interests, as seen in recent fines under the Digital Markets Act.

For Europeans the Act’s core purpose is to make technology safer and protect fundamental rights, even if it means advanced models launch a few weeks later than elsewhere. The EU’s enforcement priorities will set a benchmark worldwide—the so‑called “Brussels effect”—and must balance ethics concerns with existential risks. “Enforcement should not be headline‑driven, but should address the full spectrum of risks and ask whether fundamental rights and societal risks have been adequately considered,” says Laura Lazaro Cabrera.