EU Delays High-Risk AI Rules to 2027 and 2028 as August 2026 Model Enforcement Stands
Updated
Updated · startupfortune.com · May 19
EU Delays High-Risk AI Rules to 2027 and 2028 as August 2026 Model Enforcement Stands
7 articles · Updated · startupfortune.com · May 19
Brussels' May 7 provisional deal would push EU AI Act high-risk obligations to Dec. 2, 2027 for standalone systems and Aug. 2, 2028 for AI embedded in regulated products.
The delay does not change the Aug. 2, 2026 enforcement date for general-purpose model providers, who still face requirements on technical documentation, copyright policies, training-data summaries and cooperation with the AI Office.
Startups using third-party foundation models may avoid provider duties for the models themselves, but can still be treated as providers or deployers of high-risk systems in areas such as hiring, education, credit and essential services.
Penalties remain a major compliance driver: prohibited AI practices can draw fines of up to €35 million or 7% of global annual turnover, and the rules can apply to non-EU companies whose systems or outputs reach the EU market.
A separate transparency milestone is also emerging, with certain AI-generated content obligations now pointing to Dec. 2, 2026, leaving companies with a staggered EU compliance calendar rather than one August deadline.
Can the EU's new ban win the global arms race against AI that generates rapidly evolving abusive content?
With sexual deepfakes banned, is the world prepared for AI's next wave of financial and political fraud?
The EU’s €35 Million Crackdown: AI Nudifier App Ban and the Future of AI Compliance
Overview
The European Union has taken a strong stance against the misuse of artificial intelligence by explicitly banning AI 'nudifier' applications under the new EU AI Act. This ban, recently formalized, targets any AI system that can create or alter images to falsely portray individuals in sexually explicit or intimate contexts, making such actions illegal. The rules are classified as 'prohibited practices,' directly impacting developers and providers of these AI systems. Companies that violate these provisions face severe penalties, including fines of up to €35 million or 7% of their global annual turnover, highlighting the EU's commitment to protecting individuals from non-consensual intimate imagery.