A man was recently arrested for a burglary in a city he’d never even visited, all thanks to new AI facial recognition tech police are using. Alvi Choudhury, a 26-year-old software engineer from Southampton, was mistakenly identified as a suspect in a Milton Keynes burglary, a place over 100 miles from his home. What’s worse is that this isn’t his first brush with a false arrest.
According to LADBible, back in 2021, Alvi was arrested after being assaulted during a night out, and while his DNA was eventually removed from police records, his mugshot wasn’t. That old photo was then fed into new AI facial recognition software, leading to his recent wrongful arrest.
Alvi said he was working remotely, chatting with a client, when he heard a knock. He went downstairs, opened the door, and saw two police officers. He even joked, “Hello, officers, I’m not in trouble, am I?” with a smile.
Nothing like a dystopian error to add to your in-tray
The officer asked for his name before informing him he was under arrest for a burglary in Milton Keynes from last December. Alvi was completely floored. He said he had “never been more shocked in my life.” The officers arrested him and searched his home and phones even though he had “an abundance of evidence” showing he was at work on the day the alleged crime occurred.
Despite having a solid alibi, Alvi was held in custody for a staggering 11 hours before anyone even bothered to speak with him or look at his evidence. When he finally saw the CCTV footage that supposedly identified him, he immediately knew it wasn’t him. He said it “didn’t look like him at all.” He claims an officer even laughed when he pointed out the discrepancies.
Alvi noted that the suspect in the footage looked about 10 years younger, had lighter skin, a bigger nose, no facial hair, and different eyes and lips. He suspects that because he is a brown person with curly hair, the investigative officer simply jumped to conclusions.
The racial bias of AI-assisted technology, particularly facial recognition software, is a well-documented issue. LADBible claims the rate of false positives for Black and Asian faces is significantly higher than for white faces. For instance, Black faces experience a 5.5% false positive rate, and Asian faces see 4%, while white faces are misidentified at a mere 0.04%.
The Thames Valley Police did issue a statement, apologizing for the distress caused to Alvi. They stated that his arrest was based on the investigating officers’ “own visual assessment” that he matched the suspect in the CCTV footage, following a retrospective facial recognition match. They emphasized that the decision “was not influenced by racial profiling” and that the technology “did initially provide intelligence, but did not determine the arrest.”
While later inquiries did eliminate Alvi from the investigation, the police officers defended their actions arguing that the arrest itself wasn’t unlawful. They added that they “continue to use policing tools responsibly while striving to improve and build trust in our communities.”
Published: Feb 26, 2026 05:30 pm