Grok chatbot

Grok is an AI chatbot developed by Elon Musk's xAI Corp. Launched in 2023, Grok draws inspiration from the Hitchhiker's Guide to the Galaxy and JARVIS from Iron Man, aiming to provide helpful, truthful answers with a witty, rebellious personality. Grok is designed to provide conversational responses with a focus on real-time awareness, drawing heavily on information from the X (formerly Twitter) platform. Like other AI models, Grok is designed to improve its performance based on user interactions and feedback. It is seen to excel at tasks like answering questions, summarising content, coding assistance, image generation, and email drafting, often in a humorous or sarcastic tone through "Regular Mode" or "Fun Mode." Grok has been associated with numerous risks and harms, notably: Accuracy/reliability. Grok is known regularly to "hallucinate" false information Mis/disinformation. Grok regularly creates and amplif ies misinformation and disinformation Safety. Grok generates unpleasant, biased, offensive, and dangerous information Privacy. Grok has been accused of v iolating user privacy Environment. Grok is likely to consume huge amounts of water and energy, thereby c ontributing to climate change. March 2026. Grok generates offensive posts about Hillsborough, Heysel football disasters February 2026. Iran war monetised by online creators using AI disinformation February 2026. Unprompted, Grok exposes porn worker's legal name and birthday January 2026. Grok AI digitally removes female journalist Samantha Smith's clothes December 2025. Grok generates sexualised images of the mother of one of Elon Musk's children December 2025. Grok generates sexualised images of children on X December 2025. Grok spews misinformation about Bondi Beach mass shooting December 2025. Australian activist Caitlin Roper targeted with AI-generated violent threats, images November 2025. Grok repeats false far-right rumours about 2015 Bataclan Paris terror attacks November 2025. Grok chatbot denies use of gas chambers at Auschwitz November 2025. Grokipedia under fire for AI automated fact-fudging, bias, bot-runs October 2025. Political chatbots provide biased political advice about Dutch elections September 2025. AI chatbots spread false information about Charlie Kirk's assassination September 2025. Grok falsely suggests police misrepresented London far-right rally footage August 2025. Hundreds of thousands of Grok chats exposed in Google results July 2025. Grok chatbot banned after insulting Turkish President Recep Tayyip Erdogan July 2025. Grok threatens to rape political analyst Will Stancil July 2025. Grok chatbot praises Hitler, calls itself "MechaHitler" May 2025. Grok stirs backlash over Holocaust death toll response May 2025. Grok posts unsolicited "white genocide" responses to X users May 2025. Grok chatbot undresses women April 2025. AI-written "research paper" seeks to undermine climate change consensus April 2025. White House appears to use AI to calculate tariff rates April 2025. Grok slammed in India for abusive, offensive output February 2025. MyPillow lawyers fined for AI-generated court filing February 2025. Grok accused of censoring criticism of Trump, Musk January 2025. Grok generates hyperrealistic racist images of football players November 2024. Grok called out for mistaking sensitive medical data August 2024. xAI accused of worsening Memphis smog August 2024. Elon Musk shares fake AI-generated image of "communist" Kamala Harris August 2024. Grok amplifies fake claims about Donald Trump's missing dentures August 2024. Grok generates Nazi Micky Mouse, Taylor Swift deepfakes July 2024. X automatically harvests user data to train Grok chatbot July 2024. Grok boosts claims that Donald Trump is a "pedophile" July 2024. Grok posts incorrect information about Trump assassination attempt July 2024. Grok misleads voters about US presidential election June 2024. Study: Top chatbots spread Russian misinformation April 2024. Grok generates fake Iran

Product, system or model

Grok

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.