On June 10, 2026, the European Union took a significant step in the field of artificial intelligence (AI) by implementing new regulations that will significantly enhance its regulatory capabilities. The AI Act, a pioneering law adopted two years earlier, provides the EU with the necessary tools to oversee a rapidly evolving sector, ban certain dubious practices, and ensure transparency regarding the use of AI models. This initiative aims to protect citizens from abuses of potentially harmful technologies while fostering innovation.
A robust regulatory framework for AI
The AI Act is defined as the most advanced regulatory framework in the world regarding artificial intelligence. In addition to banning certain practices such as manipulating individuals or exploiting vulnerabilities, this law also ensures the protection of biometric data and regulates facial recognition. Companies operating in this sector will need to demonstrate that they assess and mitigate the risks associated with the technologies they develop.
Enhanced transparency obligations
With the implementation of the new provisions starting Sunday, content creators using AI will be required to clearly identify content generated by their models. The infamous deepfakes, feared for their potential to deceive the public, will be mandatory to carry appropriate labels. A grace period will be in place until December 2 for models already on the market, allowing them to comply with these new standards.
Expanded evaluation powers
The operation of the AI Office, recently established within the European Commission, marks a turning point in AI regulation. From now on, this body has the power to conduct in-depth assessments of AI models, ensuring compliance with established rules. This may include access to models before they are brought to market, providing the EU with unprecedented control over sophisticated technologies.
Inspection and sanctions
National regulators will play a key role in enforcing the rules for small-scale models. In case of non-compliance, Brussels authorities have a wide range of sanctions at their disposal. Fines can reach up to 7% of the global turnover of companies in the event of major violations. These measures aim to deter wrongful behavior and ensure that companies are held accountable.
Anticipating future technological challenges
Beyond the rules currently in place, several additional provisions will come into effect in the coming years. For example, starting in December, AIs generating inappropriate content, such as non-consensual nude images, will be completely banned. Moreover, specific regulations concerning high-risk AIs, operating in sensitive fields like health or security, will be enforced in 2027 and 2028, allowing companies to adequately prepare.
These recent developments are a crucial response for the EU in the face of increasingly advanced AI models, especially those developed outside its borders. The limited access to cutting-edge technologies has highlighted the need for effective regulation and robust oversight.
For more information on related topics, feel free to check out these articles: 300 French newspapers join forces, AI and critical thinking, Launch of a large-scale AI platform in Pays de la Loire, ALICE project for autism, and Examples of AI in public services.
