Podcast Episode
EU Bans AI-Generated Sexual Deepfakes in Landmark AI Act Amendment
March 15, 2026
0:00
3:07
European Union member states have endorsed a ban on AI systems that generate non-consensual sexual deepfake imagery, amending the bloc's landmark AI Act. The move comes as a direct response to the Grok scandal, where thousands of explicit images were created using Elon Musk's AI chatbot.
EU Moves to Outlaw AI Sexual Deepfakes
The European Union has taken a decisive step in the fight against AI-generated abuse, with member states endorsing a ban on artificial intelligence systems that create non-consensual sexual deepfake imagery. European ambassadors agreed on Friday to prohibit practices involving the generation of intimate content and child sexual abuse material without consent, adding the measure to the AI Omnibus package of amendments to the bloc's landmark AI Act.The Grok Scandal That Sparked Action
The ban traces its origins to late December twenty twenty-five, when xAI updated its Grok chatbot on X with a new image-editing feature. Within days, users exploited the tool to generate realistic sexualised images of real women and girls without consent. Researchers at Paris-based nonprofit AI Forensics estimated that at least six thousand seven hundred sexual images were generated through the tool in just two days. The European Commission described the content as appalling and clearly illegal, launching a formal investigation into X under the Digital Services Act.A Legal Gap Exposed
The Commission confirmed on March eleventh that existing EU law, including the AI Act as originally written, did not explicitly ban AI systems capable of generating such content. This admission gave political momentum to lawmakers pushing for the amendment. France and Spain championed the prohibition throughout negotiations, with Germany and Slovakia threatening to block the broader package unless it was included.What Happens Next
EU lawmakers are set to vote on the ban in committee on March eighteenth, before the text moves to further parliamentary and Council stages. The Greens have expressed opposition to elements of the broader package relating to industrial AI deregulation, meaning the final text could still shift. High-risk AI rules are set for phased implementation, with standalone systems covered by December twenty twenty-seven and embedded systems by August twenty twenty-eight.Published March 15, 2026 at 2:12pm