Beginning this Sunday, the European Union will gain the authority to enforce its landmark artificial intelligence regulations, marking a significant step in the implementation of the bloc's comprehensive digital rulebook. Following the adoption of the "AI Act" in 2024, the EU has been rolling out its governance framework in stages, utilizing a risk-based approach that scales obligations based on the potential threat to fundamental rights, health, or safety.
Under these new measures, specific AI applications are strictly prohibited, including systems designed for predictive policing, emotion recognition in schools or workplaces, and tools that manipulate human behavior. Furthermore, regulations taking effect this Sunday mandate that AI-generated content, such as deepfakes, must be clearly labeled.
Companies operating chatbots or similar systems are also required to ensure users are explicitly aware that they are interacting with artificial intelligence.
To oversee these standards, a specialized AI Office, composed of tech experts, lawyers, and economists, has been established within the EU executive. This regulatory body is empowered to test advanced, general-purpose AI models to ensure compliance. In certain instances, providers will be required to grant the EU access to their models before they are even released to the market.
If concerns remain unaddressed, the EU holds the authority to restrict a model's deployment entirely. This proactive oversight is intended to prevent past difficulties, such as the challenges the EU faced earlier this year in accessing the Mythos model from American developer Anthropic.
While national authorities will manage enforcement for smaller AI systems, the EU is prepared to penalize companies that breach the law. If a firm fails to remedy a violation, it faces significant financial penalties. The most severe fines for banned practices can reach up to seven percent of a company’s annual worldwide turnover or €35 million, whichever is higher.
Other violations or a failure to cooperate with investigations may result in fines of up to three percent of global annual revenue or €15 million.
The regulatory rollout is ongoing, with additional rules expected later this year and beyond. A ban on sexualized deepfakes is scheduled for December, a move spurred by global outcry regarding non-consensual imagery generated by platforms like Elon Musk's Grok.
However, the EU has granted some extensions for high-risk systems to allow firms time to adapt; regulations for stand-alone high-risk AI will apply from December 2027, while those embedded in other products will follow in August 2028.
Existing AI systems generally have until December 2 to comply with the current rules, with specific exemptions provided for artistic, creative, satirical, and fictional works.
The higher the risk to Europeans' rights or health, for example, the greater the obligations of firms to protect individuals from harm.
