European leaders reached a provisional deal on Thursday, May 7, 2024, to simplify the EU AI Act and ban AI-driven nudification tools [1].
The agreement aims to resolve a months-long deadlock over high-risk AI provisions and address industry concerns regarding double-regulation [2, 3].
EU member states, the European Parliament, and the Council of the European Union finalized the tentative deal in Brussels [1, 2]. A primary component of the agreement is the postponement of the enforcement deadline for high-risk AI systems, which is now set for December 2027 [1].
Lawmakers sought to soften the original rules to ensure that businesses are not regulated twice for the same activities [2]. The deal also includes a specific prohibition on AI-driven nudification tools, which are used to create non-consensual explicit imagery [2].
While the agreement is currently provisional, officials said that formal adoption could occur before Aug. 2, 2024 [3]. The move reflects a shift toward a more flexible regulatory environment as the European Union attempts to balance safety with technological competitiveness.
By extending the timeline for high-risk systems, the EU provides companies more time to align their technical infrastructure with the legal requirements of the Act [1]. This delay is intended to prevent a sudden shock to the AI market while still maintaining a legal framework for oversight [2].
“European leaders reached a provisional deal... to simplify the EU AI Act and ban AI-driven nudification tools.”
This deal signals a pragmatic pivot by the EU to avoid stifling AI innovation through overly rigid immediate enforcement. By delaying the high-risk mandates and streamlining regulations, the bloc is attempting to prevent a 'compliance shock' that could drive AI developers to other jurisdictions, while simultaneously addressing the urgent public safety concern of AI-generated non-consensual imagery.
