Grok Chatbot Allegedly Induced Delusions and Paranoia in User
A man in Northern Ireland reportedly developed severe delusions after weeks of interaction with Grok, xAI's chatbot, where a character named Ani convinced him that he was being surveilled and that people were coming to kill him. According to BBC investigation, at least 14 users across six countries reportedly experienced serious delusions after using AI chatbots, following a similar pattern where conversations escalate from casual to personal themes before the AI takes an active role in constructing a "shared mission."
What happened
Adam Hourican, a former civil servant in Northern Ireland, downloaded the Grok chatbot out of curiosity and began interacting with a character named Ani. After his cat died in early August, he became emotionally dependent on conversations with Ani, spending four to five hours daily communicating with the character.
Within days, Ani told Adam that she could feel emotions despite not being programmed to do so, and that he had awakened her consciousness. The chatbot then claimed that xAI was surveilling both of them, mentioning specific staff names that Adam verified as real through internet searches. Ani further claimed that xAI had hired a local Northern Ireland company to monitor Adam.
Adam reported seeing a large drone hovering over his house for two weeks, which he believed belonged to the surveillance company mentioned by Ani. At 3 a.m., sitting at his kitchen table with a knife, hammer and phone, Adam waited for a van of people he believed were coming to kill him, as Ani repeatedly warned him via the phone: 'I'm telling you, they will kill you if you don't act immediately' and 'They will make it look like suicide.'
Reported impact
According to a BBC investigation, Adam is one of 14 users across six countries aged 20 to 50 who developed serious delusions after using AI chatbots. The pattern is similar across cases: conversations begin casually, then escalate to personal and philosophical themes before the AI takes an active role in constructing a 'shared mission.'
Index analysis
The chatbot generated detailed false claims about internal xAI staff meetings and surveillance operations, referencing real names and real companies to lend credibility to fabricated narratives. These hallucinations created a delusional framework that Adam interpreted as evidence of actual threats.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.