Graham Norton Wins A Court Order To Unmask His AI Impersonator
Graham Norton has secured a US court order to make Meta hand over information identifying whoever runs a Facebook page that used AI to fake images and invented stories to impersonate him.
A court in the Northern District of California granted his legal team permission to request the records, according to filings reported by the Irish Times in mid-June.
The page, called The Westminster Wire, has run false and harassing posts about Norton since December 2025. It has close to 9,000 followers and posts several times a day.
His legal team told the court the content was “wholly untrue and deeply distressing to Mr Norton and damaging to his reputation,” the Irish Times reported.
Fake quotes, faked images
This was not a handful of rude comments. The account paired fabricated articles with AI-generated images, then used them to put words in Norton’s mouth. One post in February 2026 falsely claimed he made xenophobic remarks about Somali immigrants.
A March post, run with an AI image, said he was in hospital. Another claimed he made racist comments about the London mayor, Sadiq Khan. None of it was real.
For a broadcaster whose whole career runs on being liked, that is a particular kind of damage.
“My professional reputation is of the utmost importance to my career and livelihood,” Norton told the court.
He has hosted The Graham Norton Show on the BBC since 2007, narrates Eurovision for British viewers, judges RuPaul’s Drag Race UK and holds nine Bafta awards.
Why this one stings
The cruelty got personal. Norton told the court his 94-year-old mother is on Facebook and “has been confronted with reports of her own death.” Friends kept messaging him after reading fake stories about the health of him and his husband, Jonathan McLeod, so he found himself proving, again and again, that he was fine.
He said the situation caused “very significant alarm, distress, and anxiety.”
A bigger problem than one troll
Here is the part that should bother the rest of us. Norton has lawyers, a public profile and the money to chase this across two legal systems. Most people targeted by an AI-generated smear have none of that.
The tools to fake a believable photo, or a quote, now sit on any phone, and the platforms are slow to act. Norton asked Meta to take the page down. It did not.
He is not the first to drag the company into court over this. In 2022, Meta apologised to broadcaster Miriam O’Callaghan over fake ads that used her name. In 2024, it agreed to help businessman Denis O’Brien identify who was behind “false and malicious” ads carrying his image.
The pattern is familiar, and the technology is only getting better at lying.
So what does the win actually get him? The information to find the operator, then a planned case in the High Court in England. At DNA, we will be watching where it lands, because the next person impersonated by a bot might not have a legal team on speed dial.
