Brief IA

Europe: Strategic Delay in AI Regulation Until 2028

⚖️ Regulation & Ethics·Tom Levy·

Europe: Strategic Delay in AI Regulation Until 2028

Europe: Strategic Delay in AI Regulation Until 2028
Key Takeaways
1Europe postpones the implementation of AI regulations in critical sectors such as biometrics and education to December 2027.
2AI products, such as toys, will see their regulations applied starting in August 2028, according to the new European agreement.
3SMEs will benefit from exemptions on documentation and easier access to regulatory sandboxes.
💡Why it mattersThis delay could impact Europe's competitiveness in light of the rapid advancements in AI worldwide.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

Europe Chooses Delay for AI Regulation

The European Commission, the European Parliament, and the EU Council have decided to simplify the rules surrounding artificial intelligence, but by postponing their implementation. This agreement is based on the "Digital Omnibus on AI," a legislative package that modifies the current AI law.

The rules regarding AI systems in areas such as biometrics, critical infrastructure, education, and migration will only take effect from December 2027. For AI products like toys and "elevators," the application is pushed back to August 2028. Furthermore, systems generating sexually explicit content without consent, including "nudification" applications, are now explicitly prohibited.

Relief for Small and Medium Enterprises

Small and medium enterprises, defined as those with up to 750 employees and a turnover of 150 million euros, will benefit from relief measures. This includes reduced registration and documentation requirements, as well as better access to "regulatory sandboxes," testing environments that allow for experimentation with AI under real conditions.

Labeling and Other Obligations

Labeling requirements will come into effect in August 2026. Joerg Heidrich, a lawyer specializing in information technology law, notes that the labeling obligations under Article 50 of the AI regulation remain largely unchanged. Companies will need to label deepfakes and certain AI-generated texts starting August 2, 2026. However, this rule only applies to fully automated content that has not been reviewed by a human, thereby limiting its impact.

A Framework Awaiting Approval

This proposal, introduced in November 2025 as part of the EU's simplification agenda, still needs to be formally approved by the Parliament and the Council. The goal is to enhance Europe's competitiveness while ensuring the protection of its citizens.

The European Commission describes the outcome as "favorable to innovation," but Joerg Heidrich, legal advisor at Heise, expresses reservations. He believes that the postponement of deadlines for high-risk categories to December 2027 or August 2028 was inevitable, as the initial timeline of August 2026 was practically impossible to meet.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.