EU AI Act Amendment: European Council Proposes Ban on Nudification Tools

European Council Proposes Ban on AI Nudification Tools

The European Council has proposed amendments to Regulation (EU) 2024/1689, the EU AI Act, that explicitly ban AI nudification tools used to generate non-consensual sexual and intimate content or child sexual abuse material. This follows a similar proposal from the European Parliament, indicating the measure is likely to be included in the final negotiated plan. The amendments also reinstate stricter standards for processing special categories of personal data, requiring 'strict necessity' for bias detection and correction purposes, and mandate that providers register AI systems in an EU database if they claim exemptions from high-risk rules. These changes respond to incidents like the Grok chatbot generating millions of non-consensual intimate images, which prompted a European Commission probe.

Why This Amendment Matters for AI Governance

Nudification tools use AI to create or alter images or videos to depict individuals in sexual or intimate contexts without their consent. Under the EU AI Act, these tools are classified under prohibited AI practices (Article 5), which apply from 2 February 2025. The European Council's amendment clarifies and strengthens this ban, addressing growing concerns about AI-generated harmful content. This move aligns with the Act's risk-based approach, where unacceptable risk practices are banned outright to protect fundamental rights.

For businesses, this amendment underscores the urgency of AI governance compliance. The EU AI Act imposes severe penalties for violations of prohibited practices: up to EUR 35 million or 7% of global annual turnover. With the full applicability of the Act set for 2 August 2026 (with some exceptions until 2027), organizations developing or using AI must proactively assess their systems. This includes not only nudification tools but also other high-risk AI systems, such as those used in recruitment or critical infrastructure, which face obligations from 2 August 2026.

The broader impact extends to AI ethics and transparency. The amendment reinforces the EU's commitment to stringent AI oversight, with the EU AI Office established within the European Commission to oversee general-purpose AI models and coordinate enforcement. Companies operating in the EU must navigate these evolving requirements, which may influence global AI standards. For more on implementation, see our EU AI Act compliance roadmap guide.

Practical Steps for Organizations to Ensure Compliance

To avoid non-compliance risks, organizations should take immediate action to align with the EU AI Act's prohibited practices and broader governance rules. Here are key steps to consider:

1. Conduct an AI System Inventory: Identify all AI tools in use, including those for content generation, image processing, or data analysis. Flag any systems that could be misused for nudification or similar prohibited purposes. This inventory should be reviewed regularly as part of your AI governance framework.
2. Assess Against Prohibited Practices: Evaluate AI systems against Article 5 of the EU AI Act, which bans practices like manipulative AI, social scoring, and real-time remote biometric identification in public spaces (with exceptions). The nudification ban falls under this category. Use frameworks like the NIST AI Risk Management Framework (AI RMF 1.0) or ISO/IEC 42001 for structured assessments.
3. Implement Ethical AI Frameworks: Adopt policies that prohibit the development or use of nudification tools and other banned AI. Integrate ethical guidelines into AI development lifecycles, emphasizing consent, privacy, and human oversight. Tools like Holistic AI and OneTrust offer compliance monitoring features to help track AI risks and regulatory changes.
4. Stay Updated on Regulatory Changes: The EU AI Act is evolving, with codes of practice for general-purpose AI models expected by 2 May 2025 and governance rules applying from 2 August 2025. Subscribe to regulatory intelligence platforms like AIGovHub for real-time updates on amendments and enforcement trends. Check our best AI governance platforms review for vendor comparisons.
5. Prepare for High-Risk AI Obligations: If your AI systems fall under high-risk categories (e.g., recruitment, healthcare), ensure you meet transparency and documentation requirements by 2 August 2026. This includes conducting conformity assessments and maintaining technical documentation. For guidance, refer to our modifying AI systems compliance guide.

Organizations should also monitor related developments, such as the Colorado AI Act (effective 1 February 2026) and NYC Local Law 144 (effective 5 July 2023), which address AI bias in employment. For insights on AI governance gaps, see our AI safety incidents analysis.

Conclusion: Proactive Compliance Is Key

The European Council's proposal to ban nudification tools highlights the EU's aggressive stance on AI ethics and safety. With prohibited practices applying from February 2025 and full compliance required by August 2026, businesses must act now to audit their AI systems, implement robust governance, and avoid hefty penalties. Leveraging tools like AIGovHub's AI governance platform can provide real-time regulatory alerts and vendor assessments to streamline this process. Remember, this content is for informational purposes only and does not constitute legal advice. Always verify the latest regulatory timelines with official sources.