EU bans nudifier apps, industry pushes back

The European Parliament and Council have agreed to ban AI systems used to create non-consensual intimate deepfakes, with the prohibition taking effect on December 2, 2026, according to a Parliament press release. The measure targets so-called “nudifier” apps and applies not only to companies placing such systems on the EU market, but also to deployers who use them to generate the content. (europarl.europa.eu) At the same time in Washington, the data-center industry is organizing a rebuttal to claims that the AI buildout is pushing electricity costs onto households, according to Axios. (europarl.europa.eu) The push comes as local opposition to large data centers has spread and as regulators and lawmakers examine who should pay for new power demand. ### What exactly did the EU agree to ban? The May 7 European Parliament release said the ban covers AI systems that create child sexual abuse material or depict the intimate parts of an identifiable person, or show that person engaged in sexually explicit activities, without consent. The content can be images, video or audio, the release said. (europarl.europa.eu) The same release said the prohibition applies in three places: putting such systems on the EU market for that purpose, putting them on the market without reasonable safety measures to prevent that use, and deployers using the systems to create the content. Companies have until December 2, 2026 to bring systems into line. (axios.com) ### Why is the start date December 2? December 2, 2026 appears repeatedly in the EU package because lawmakers tied several AI Act changes to that date. The Parliament release said watermarking obligations for AI-generated content were also moved to December 2, 2026, while some high-risk AI obligations were postponed to 2027 and 2028. (europarl.europa.eu) Euronews reported on May 19 that the nudifier ban was part of the bloc’s broader AI-law revisions and would cover both developers and users creating false intimate content of real people without consent. That makes the rule one of the clearest AI-specific prohibitions now scheduled in the EU framework. (europarl.europa.eu) ### What is happening in Washington on data centers? Axios reported on May 18 that tech and AI infrastructure companies are stepping up efforts to challenge the argument that consumers are being stuck with the bill for data-center growth. The report said the industry plans to take that case to Capitol Hill and to regulators. The lobbying push follows a broader backlash around large data centers. (europarl.europa.eu) CNBC reported in March that hyperscalers building AI infrastructure were facing scrutiny over whether their electricity demand could raise household power bills, and House lawmakers previously asked utilities how they would protect consumers from data-center-related costs. ### Why are power costs at the center of that fight? (msn.com) PJM Interconnection’s market has become a focal point for the debate. Bloomberg reported on May 14 that power prices on the largest U.S. grid rose 76% in the first quarter amid demand from data centers, increasing pressure on grid operators and policymakers. (axios.com) Other groups have argued the risks go beyond wholesale prices. The Union of Concerned Scientists said in a February report that unmitigated data-center growth could expose the public to higher utility bills and other costs, while local and state fights over data-center siting have already delayed or blocked projects. ### What comes next? December 2, 2026 is the next fixed date in Europe, when companies must have systems aligned with the new ban and AI-generated-content watermarking rules also begin to apply, according to the European Parliament. (cnbc.com) In the United States, the next phase is political rather than statutory: the industry’s case will be made to Congress and regulators as ratepayer protections and power-contract rules remain under review. (europarl.europa.eu) (ucs.org) (bloomberg.com)