EU Parliament Approves Key Amendments to AI Act, Adjusting Compliance Timelines and Banning Deepfake Abuse
The European Union's pioneering Artificial Intelligence Act, a global benchmark for AI regulation, recently underwent its first formal amendments following a vote in the European Parliament on June 16, 2026. The changes, part of a broader simplification initiative across the EU, primarily focus on adjusting compliance timelines for high-risk AI applications and introducing stringent prohibitions. Notably, the main compliance deadline for AI systems deployed in critical sectors such as healthcare, employment, and law enforcement has been shifted from August 2, 2026, to December 2, 2027. For AI integrated into regulated products like medical devices and industrial machinery, the deadline is extended even further, to August 2, 2028.
A significant addition to the Act is the immediate ban, effective December 2, 2026, on AI systems designed to generate nonconsensual intimate imagery or child sexual abuse material (CSAM). This prohibition applies to both developers and deployers of such tools, underscoring the EU's commitment to combat digital exploitation. Furthermore, the amendments clarify the regulatory landscape for AI-enabled industrial machinery, resolving previous overlaps with existing product safety laws and aiming to prevent redundant compliance burdens. The European Council is expected to formalize these amendments, marking a crucial step in the evolving governance of artificial intelligence.
