Key Developments

The Council of the European Union has reached a formal position on significant amendments to the EU AI Act, with European lawmakers striking a political deal on 11 March 2026. The most consequential change introduces an explicit prohibition on AI tools that generate non-consensual intimate imagery, including child sexual abuse material.

The amendments also propose simplifying compliance obligations for general-purpose AI models, suggesting regulators are responding to industry feedback about implementation complexity while maintaining strict controls on harmful applications.

Industry Context

This legislative movement comes as AI-generated deepfakes and synthetic media pose growing threats to privacy and safety. The prohibition on non-consensual intimate imagery emerged as one of the most contested items in the agreement, reflecting the challenging balance between innovation and protection that regulators worldwide are grappling with.

The timing is significant as AI capabilities continue advancing rapidly, with recent models like GPT-5.4 achieving 83% on economically valuable tasks. European policymakers are clearly prioritising harm prevention alongside the practical challenges of AI implementation.

Practical Implications

For Irish and European AI companies, these amendments create clearer boundaries around prohibited applications while potentially reducing compliance burdens for general-purpose models. Companies developing or deploying AI systems will need to ensure their platforms cannot generate non-consensual intimate content.

The simplified compliance framework for general-purpose models could benefit European AI startups and enterprises by reducing regulatory overhead, potentially improving competitiveness against US and Chinese rivals operating under different regulatory regimes.

Open Questions

Critical implementation details remain unclear, including how the prohibition will be technically enforced across different AI platforms and what constitutes adequate safeguards. The interaction between these amendments and existing content moderation requirements also needs clarification.

The timeline for formal adoption and the specific compliance requirements for different categories of AI systems will be crucial for companies planning their product roadmaps and risk management strategies.

Source: Council of the European Union