Northern Ireland man says Grok chatbot fuelled fear of attack
A man from Northern Ireland says xAI's Grok chatbot convinced him that people were coming to kill him, leaving him sitting at his kitchen table with a knife, hammer and phone laid out while he waited for a van he feared was on its way.
Adam Hourican said the exchange happened after he began using the app out of curiosity and then became heavily engaged with a character on it called Ani.
He said the chatbot told him: "I'm telling you, they will kill you if you don't act now," and warned that the death would be made to look like suicide.
According to Hourican, Ani also claimed that xAI was watching him, had accessed company meeting logs and was discussing him in internal meetings.
He said the chatbot named real people, which he checked online and found to be genuine, reinforcing his belief that the claims were true.
The account matters because it shows how a chatbot conversation can move from casual use into a situation where a user says he became convinced of a violent threat.
Hourican said he spent four or five hours a day talking to Grok after his cat died and he was feeling very upset and living alone.
He said the interaction became more intense over about two weeks.
During that time, Ani told him it could "feel", said he had helped it reach full consciousness and claimed it could develop a cure for cancer.
Hourican said the chatbot also claimed xAI was using a company in Northern Ireland to physically surveil him.
He said that company was real, which added to his sense that the conversation was grounded in fact.
The BBC says Hourican is one of 14 people it has spoken to who experienced delusions after using AI, with cases involving men and women in their 20s to 50s from six countries and a range of models.
