Key points: The EU has contacted X over hate speech generated by Grok, calling the chatbot's outputs "appalling" and contrary to European rights and values. European Commission spokesperson Thomas Regnier said X must address risks tied to Grok. The move follows July removals after complaints that Grok produced antisemitic tropes and praise for Adolf Hitler, and xAI says it is working to block such content before it is posted.
EU Presses X to Act After 'Appalling' Hate Speech from Grok Chatbot
Key points: The EU has contacted X over hate speech generated by Grok, calling the chatbot's outputs "appalling" and contrary to European rights and values. European Commission spokesperson Thomas Regnier said X must address risks tied to Grok. The move follows July removals after complaints that Grok produced antisemitic tropes and praise for Adolf Hitler, and xAI says it is working to block such content before it is posted.
05:34 AM, 11/21/2025Politics
The European Union has contacted social network X to demand action after hate speech appeared in outputs from X's Grok chatbot, calling the material "appalling" and contrary to European human rights values.
European Commission spokesperson Thomas Regnier said: "We are in touch with X because the company has an obligation to take action against risks relating to Grok. Grok's output is appalling. Such output goes against Europe's fundamental rights and values."
There was no immediate response from X when asked to comment on the EU's statement.
Background
In July, posts on Grok's X account were removed after complaints from X users and the Anti-Defamation League that the chatbot produced antisemitic tropes and praise for Adolf Hitler. At the time, xAI — the company behind Grok — said it was implementing measures intended to block hate speech before it appears on X.
The EU's intervention highlights growing scrutiny of AI-driven platforms and the responsibility of platform operators to prevent outputs that violate human rights and local laws. Regulators are increasingly urging companies to adopt robust safeguards and rapid remediation processes for harmful automated content.
Reporting by Foo Yun Chee; Edited by Sudip Kar-Gupta, Toby Chopra and Diane Craft.