EU Approves Nude App Ban, Delays AI Act
▼ Summary
– European lawmakers voted to delay compliance deadlines for high-risk AI systems in the EU AI Act until December 2027.
– They also backed a proposal to ban nudify apps, though details are unspecified and exceptions for systems with safety measures are noted.
– The vote pushes back deadlines for companies under sector-specific rules to August 2028 and delays watermarking rules to November 2026.
– This decision follows public outrage over sexualized AI deepfakes on social media earlier in the year.
– The changes create further uncertainty for businesses and require negotiation with the European Council before becoming law.
In a significant move, the European Parliament has voted to postpone major enforcement timelines for the landmark EU AI Act while simultaneously endorsing a new prohibition on so-called nudify apps. This decision creates a new phase of regulatory uncertainty for technology companies across the continent. The vote, passed with a substantial majority, directly impacts the schedule for what are classified as high-risk AI systems, pushing their compliance deadline from this summer to December 2027. These systems are defined as those presenting a serious potential threat to health, safety, or fundamental human rights.
For AI developers whose products fall under existing sectoral regulations, such as those governing medical devices or toys, the proposed timeline is even longer. Their new deadline for aligning with the AI Act’s requirements is now August 2028. Additionally, the implementation of rules mandating that providers use watermarks on AI-generated content has been deferred until November 2026. All of these provisions were initially slated to become enforceable in August of this year.
Parliamentarians also supported amendments to incorporate a ban on applications designed to create non-consensual AI-generated nude imagery. While the precise legislative language remains undefined, the proposed ban includes a critical caveat. It would reportedly not apply to AI systems with effective safety measures that technically prevent users from generating such exploitative content. This legislative push follows intense public and political backlash across Europe earlier this year, driven by the proliferation of sexualized deepfakes on social media platforms.
This parliamentary vote extends a prolonged period of ambiguity for the business community. Companies have already navigated delays after the European Union missed its own deadlines for publishing essential implementation guidance and made subsequent adjustments to the law’s framework. A further complication is the legislative process itself. The European Parliament cannot enact these changes independently; it must now enter negotiations with the European Council, which comprises ministers from all 27 member states, to agree on a final legal text. It remains uncertain whether this revised timeline can be formally adopted before the original August enforcement date arrives.
(Source: The Verge)
