EU Parliament wants to ban Nudifier apps

Developments in AI often outpace legislation. This is particularly evident with so-called Nudifier apps, which use generative AI to create realistic-looking nude images from ordinary photos in the form of deepfakes. The EU Parliament voted by an overwhelming majority on Thursday that such systems should be banned in the future if they generate sexually explicit content without the consent of the depicted individuals. 569 MEPs voted in favour, 45 against, and 23 abstained.

The initiative is part of a comprehensive positioning by MEPs on the so-called Digital Omnibus Regulation. The people's representatives aim for a practical adaptation of the AI Regulation.

To give companies more planning security, Parliament is proposing an adjustment of the deadlines in the AI Act. High-risk AI systems used in biometrics, education, or law enforcement, for example, should only be subject to the full requirements from December 2, 2027. For AI applications in the area of general product safety, August 2028 is even being discussed.

Protection against misuse and more transparency

However, Parliament remains unyielding on the digital integrity of the individual. The planned ban on de-nuding AI directly targets providers like Elon Musk's Grok, whose software is misused to digitally degrade women and children. The only exceptions to this ban would be systems with effective technical security barriers that prevent the generation of such deepfakes. In addition, MEPs are pushing for a labeling requirement for AI content by November 2026 to make its origin transparent.

With the outlined tightening measures, MEPs are reacting to a wave of digital sexual violence. Prominent cases like that of moderator Collien Fernandes illustrate the destructive power of the technology. She fell victim to such AI manipulations and was confronted with deceptively real, but fake, nude images online. This served as a warning in the parliamentary debate about how quickly technological playthings can escalate into massive blackmail and psychological violence.

Despite much agreement on these protective measures, resistance is mounting against other parts of the package. The TÜV Association, for example, warns of a looming "sectoral exit." Key high-risk areas such as medical devices, machinery, or toys are to be removed from the direct scope of the AI Regulation and instead be regulated in individual sectoral laws.

With the Parliament's vote in line with the committee's recommendation, trilogue negotiations with the Council of Ministers can now begin to determine the final shape of the law. The EU states have already taken a similar position, which should allow for a swift agreement.

(mho)