Starting Sunday, the European Union will have the authority to enforce its landmark artificial intelligence regulations, imposing fines on companies that violate the law. The EU's AI Act, adopted in 2024, introduces a risk-based approach to AI, with higher obligations for firms depending on the risk to individuals' rights or health. The new rules also ban certain AI uses, such as predictive policing and emotion recognition in workplaces or schools.
Key Provisions of the New AI Rules
From Sunday, AI-generated content like deepfakes must be clearly labelled, and firms must ensure their AI systems, such as chatbots, make it evident to users that they are AI. Existing AI systems have until December 2 to adapt to these new rules, with exemptions for artistic, creative, satirical, or fictional work.
Enforcement and Fines
The EU's AI Office, a regulatory body attached to the EU executive, will have the power to enforce the bloc's rules on the most advanced AI models. This office, composed of tech experts, lawyers, and economists, will be able to test AI models to ensure compliance with the law. Companies must grant the EU access to their models, and in some cases, the EU can demand access before they are placed on the market.
The EU can also restrict a model's deployment within the EU if concerns arise that are not addressed by the provider. National authorities will enforce the rules for smaller AI systems. Companies found breaking the rules can be fined up to seven percent of their annual worldwide turnover or 35 million euros, whichever is higher, for the most severe violations.
Future Implementations and Delays
More rules under the AI Act will be enforceable later this year and beyond. From December, there will be a ban on AI systems generating sexualized deepfakes. The EU is delaying the implementation of some high-risk AI rules to give companies more time to prepare, with these rules applying to stand-alone high-risk AI systems from December 2027 and embedded AI tools in August 2028.
