Ex-Officer in Hiding After AI Falsely Links Her to Murder Case
Ex-Officer in Hiding After AI Falsely Links Her to Murder Case

A former police officer has been forced to move to a safe house after artificial intelligence and social media users falsely identified her as one of the arresting officers in the Henry Nowak murder case. Christi Hill, who left Hampshire Constabulary in April 2024, 20 months before the incident, said she feared for her safety after the misinformation spread online.

The confusion arose when Grok, the AI service on X (formerly Twitter), wrongly linked Hill to the arrest of Henry Nowak in December 2025. Nowak had been attacked by Vickrum Digwa, who was later jailed for life after lying that the 18-year-old had racially abused him. Hampshire Constabulary faced criticism for its response, with an officer heard telling the dying Nowak, “Don’t think you have, mate,” when he said he had been stabbed.

Hill wrote on social media: “Today, my name and image have been widely circulated on social media, and now by AI platforms such as Grok, falsely identifying me as one of the arresting officers in the Henry Nowak case. To be absolutely clear, I was not involved in this incident.” She said the misidentification stemmed from a national police bravery award media release featuring her photo.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Her mother, Nikki Hill, told The News in Portsmouth: “It's been terrible seeing her name and face all over the internet calling her a murderer.” Hill added that she had received little support from Hampshire Constabulary in correcting the false narrative.

Hampshire Constabulary said in a statement: “A police officer unrelated to this case has been misidentified online and subject to death threats. We ask that people avoid harmful speculation online.” The Independent Office for Police Conduct is investigating the force’s response on the night of the incident.

Grok posted an apology on X, stating: “This mistake fueled harmful online speculation, threats, and real distress, including forcing her into a safe house. I take responsibility. AI systems can err on identifications, especially with limited public details and unverified reports.”

Pickt after-article banner — collaborative shopping lists app with family illustration