Skip to main content
NOPE
Medium reported severity

Southampton: Alvi Choudhury arrested at home on suspicion of a Milton Keynes burglary after a police retrospective facial recognition match, held nearly 10 hours, claiming damages

The Guardian, in a joint report with Liberty Investigates, reported on 25 February 2026 that Alvi Choudhury, a 26-year-old software engineer, was arrested at his home in Southampton in January 2026 on suspicion of a £3,000 burglary in Milton Keynes, about 100 miles away, and held in custody for nearly 10 hours. Thames Valley Police had used automated retrospective facial recognition software, which matched him with CCTV footage of the burglary suspect. Choudhury says the man in the footage looked about 10 years younger and had different features. Thames Valley Police wrote to him that the arrest may have been the result of bias within facial recognition technology, and told the Guardian that the decision rested on the investigating officers’ own visual assessment and was not influenced by racial profiling. Choudhury is claiming damages from Thames Valley Police and Hampshire Constabulary.

AI system
Unidentified facial recognition system
Not stated in the sources for this search (the Guardian says UK police forces use an algorithm procured by the Home Office from Cognitec)
Occurred
8 Jan 2026
Reported
25 February 2026
Event location
United Kingdom
What the AI did
Made a claim about the person
Reported harm
Loss of LibertyPsychological Distress
Whose AI use
An institution’s AI use
Setting
Justice · Privacy
Evidence
AI involvement supported · Causal attribution disputed · 4 sources, 3 underlying accounts
10 claims: 1 documented, 1 disputed, 8 reported. 6 open questions
People reported harmed
1 person

AI system as recorded: Retrospective facial recognition search run by Thames Valley Police (product, version and reference database not stated in the sources)

Reported harm

The Guardian reports Choudhury was handcuffed at home, held for nearly 10 hours and released at 2am, and was unable to work the following day. Choudhury says neighbours saw him led away. Thames Valley Police apologised for the distress caused and deny the arrest was unlawful.

What remains unknown

  • Which product, version, face-match threshold and match score the Thames Valley Police search used is not stated. The National Physical Laboratory report evaluates Cognitec FaceVACS-DBScan ID v5.5 (the Police National Database software) at stated thresholds, and the sources do not say those settings applied here.
  • The documents shared with the Guardian, the police letter to Choudhury, the CCTV footage and the police interview record were not inspected. All account details come from news reports and a chambers post.
  • Whether the human visual assessment preceded the arrest request, which force made the request and which force made the decision are stated differently by Choudhury, his solicitors and the police, and no record was inspected.
  • Whether a court claim was filed and the outcome of the complaint and the damages claim are not reported in any inspected source (the latest inspected reports are dated 1 April 2026).
  • The date 8 January 2026 comes from the news post of the claimant’s counsel (Internet Archive capture). The Guardian says only January. Distances given by outlets differ (100 miles in the Guardian) and are not material.
  • The burglary, the burglary suspect and any prosecution are not described in the inspected sources.

What the evidence supports

AI involvement: supported. Thames Valley Police’s own statement, quoted by the Guardian, refers to a retrospective facial recognition match preceding the arrest, and the Guardian cites documents shared by Liberty Investigates. The police say the arrest decision rested on the investigating officers’ own visual assessment. Choudhury says the footage showed a man who looked about 10 years younger with different features, and that he assumed the investigating officer decided to arrest him because he is a brown person with curly hair. The police also wrote that the arrest may have been the result of bias within the technology. The weight of the match in the arrest decision is disputed. The sources do not identify the product version, threshold or match score.

10 claims: 1 documented, 1 disputed, 8 reported. What the statuses mean

Reported Alvi Choudhury was arrested at the home he shares with his parents in Southampton on 8 January 2026 on suspicion of a burglary in Milton Keynes, handcuffed and held in custody for nearly 10 hours until released at 2am. Hampshire Constabulary officers carried out the arrest.

Causal attribution. Event description only. The account comes from Choudhury and his solicitors through the Guardian, Liberty Investigates and a chambers news post. The Thames Valley Police statement quoted by the Guardian treats the arrest as having happened and apologises for the distress caused. The 8 January date appears only in the chambers post (the Guardian says January).

Reported Thames Valley Police used automated retrospective facial recognition software that matched Choudhury with CCTV footage of a suspect in the Milton Keynes burglary. Choudhury says he learned about a week after the arrest, after filing a complaint, that an automated facial recognition system had flagged him as a possible suspect.

Causal attribution. The Guardian bases the match on documents shared by Liberty Investigates, which were not inspected. The police spokesperson statement quoted in the same report itself refers to a retrospective facial recognition match. The sources do not state which product, version, threshold or match score applied to this search.

  • theguardian.com(opens in new tab) supports · English
    Thames Valley police had used automated facial recognition software which matched him with footage of a suspect of a £3,000 burglary 100 miles away in Milton Keynes
  • theguardian.com(opens in new tab) supports · English
    following a retrospective facial recognition match, and was not influenced by racial profiling.
  • libertyinvestigates.org.uk(opens in new tab) supports · English
    A week later, after filing a complaint against Thames Valley Police, he found out the full story: he’d been flagged as a possible suspect by an automated facial recognition system.
  • theguardian.com(opens in new tab) context · English
    UK police forces use an algorithm procured by the Home Office from Cognitec, a German company.
Reported Thames Valley Police told the Guardian that the decision to arrest was made after a human visual assessment as well as the match, denied that the arrest was unlawful and said it was not influenced by racial profiling, and apologised for the distress caused. The Guardian reports the National Police Chiefs’ Council position that facial matches should be treated as intelligence, not fact.

Causal attribution. A party statement about its own conduct, relayed by the Guardian. It addresses the arrest decision and gives no detail of the match.

  • theguardian.com(opens in new tab) supports · English
    Thames Valley police said the decision to arrest Choudhury was made after a human visual assessment as well.
  • theguardian.com(opens in new tab) supports · English
    A Thames Valley police spokesperson denied the arrest was unlawful and said: “While we apologise for the distress caused to the complainant in this case
  • theguardian.com(opens in new tab) supports · English
    Facial matches should be treated as intelligence, not fact, according to the National Police Chiefs’ Council.
Reported Choudhury says the man in the CCTV footage looked about 10 years younger than Choudhury and had different features, that he offered evidence of work meetings in Southampton on the day of the burglary, that officers at the Hampshire police station laughed when he asked whether the footage looked like him, and that the Thames Valley Police officers who came to interview him said they knew he was not the suspect after looking at the footage and his picture.

Causal attribution. First-person account relayed by the Guardian. The CCTV footage and the police interview record were not inspected, and no source independent of Choudhury describes the suspect’s appearance or the interview.

  • theguardian.com(opens in new tab) supports · English
    “I was very angry, because the kid looked about 10 years younger than me,” said Choudhury, who wears a beard.
  • theguardian.com(opens in new tab) supports · English
    He offered evidence of work meetings in Southampton on the day of the crime but he was instead taken into custody.
  • theguardian.com(opens in new tab) supports · English
    But Choudhury said officers at the Hampshire police station laughed when he asked: “Does this look anything like me?”
  • theguardian.com(opens in new tab) supports · English
    And he said the Thames Valley police officers who arrived to interview him said “they knew I wasn’t the suspect after looking at footage of the suspect and looking at my picture”.
Reported Thames Valley Police wrote to Choudhury that the arrest “may have been the result of bias within facial recognition technology”. An officer also told him that, because facial recognition use is already subject to review at a strategic level, the issue would not be raised as part of wider organisational learning.

Causal attribution. Quotations from a police letter to Choudhury as relayed by the Guardian and Liberty Investigates. The letter itself was not inspected. The wording is conditional (“may have been”).

  • theguardian.com(opens in new tab) supports · English
    Thames Valley police admitted to Choudhury the arrest “may have been the result of bias within facial recognition technology”.
  • libertyinvestigates.org.uk(opens in new tab) supports · English
    Thames Valley Police wrote to Choudhury, acknowledging that his arrest “may have been the result of bias within facial recognition technology”
  • theguardian.com(opens in new tab) supports · English
    as the use of facial recognition is already subject to review at a strategic level, I do not feel the need to raise this issue as part of wider organisational learning
Disputed Whether the facial recognition match or the officers’ own visual assessment led to the arrest decision, and whether bias in the software contributed, is disputed. Choudhury says he assumed the investigating officer saw a brown person with curly hair and decided to arrest him. Thames Valley Police say the arrest rested on the investigating officers’ own visual assessment following the match and was not influenced by racial profiling, while also acknowledging in writing that the arrest may have been the result of bias within facial recognition technology.

Causal attribution. Two accounts that conflict on the decisive input to the arrest. Neither the match record nor the officers’ assessment record was inspected.

  • theguardian.com(opens in new tab) supports · English
    “I just assumed that the investigative officer saw that I was a brown person with curly hair and decided to arrest me.”
  • theguardian.com(opens in new tab) supports · English
    their arrest was based on the investigating officers’ own visual assessment that the individual matched the suspect in CCTV footage
  • libertyinvestigates.org.uk(opens in new tab) supports · English
    Confusingly, however, the force concluded that his arrest was not a case of racial profiling, because officers had made their own assessment of the images before arresting him.
Reported Choudhury reports that neighbours saw him led away in handcuffs, that his father was very anxious about him being held, and that he was unable to work the following day. He says he sometimes needs security clearance to work for government clients, is asked about arrests, and that “This makes me look dodgier and dodgier”, and that he is afraid the automated system could trigger more arrests now that a second mugshot has been taken.

Causal attribution. First-person account relayed by the Guardian. The police apology for distress caused is the only independent acknowledgement of an effect. No effect on an actual clearance decision is reported.

  • theguardian.com(opens in new tab) supports · English
    His neighbours saw him being led away in handcuffs, his father was very anxious about him being held and he was unable to work the following day, he said.
  • theguardian.com(opens in new tab) supports · English
    He sometimes needs security clearance to work for government clients and he is asked about arrests and said: “This makes me look dodgier and dodgier.”
  • theguardian.com(opens in new tab) supports · English
    Now he has had a second mugshot taken he is afraid the automated system could trigger more wrongful arrests.
  • theguardian.com(opens in new tab) supports · English
    While we apologise for the distress caused to the complainant in this case
Reported Choudhury’s image was held on the police system only because of an arrest in 2021 that the Guardian describes as wrongful (he had been attacked on a night out) and that ended with police releasing him with no further action.

Causal attribution. Background as told by Choudhury to the Guardian. The sources do not state in terms that this custody image was the one returned by the search.

Reported Choudhury filed a complaint against Thames Valley Police and is claiming damages from Thames Valley Police and Hampshire Constabulary. His solicitors have instructed counsel to pursue a claim for damages against Thames Valley Police for false imprisonment, breach of data protection law, breach of Article 8 ECHR and breach of the Equality Act 2010.

Causal attribution. A complaint and a claim establish only their own existence and the allegations made. No court filing or outcome was found in the inspected sources.

Documented A National Physical Laboratory report for the Home Office and the Office of the Policing Chief Scientific Adviser (dated October 2025) evaluated Cognitec FaceVACS-DBScan ID v5.5, which it describes as the facial recognition software used in the Police National Database for retrospective facial recognition. At a face-match threshold of 0.8 on its test data it found a false positive identification rate of 4.0 % for Asian subjects and 0.04 % for White subjects.

Causal attribution. The report establishes the evaluation findings on its own test dataset. It does not state which product version, threshold or settings applied to the search in this case.

Sources

4 sources inspected, from 3 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read through an Internet Archive capture of 22 April 2026 because the site returned HTTP 403 to direct requests. Applies to s3.

Event countries: United Kingdom. Affected-person countries: United Kingdom. Court countries: Unknown.

The arrest took place at Choudhury’s home in Southampton and the burglary was in Milton Keynes, both placed in England by the Guardian (Thames Valley Police and Hampshire Constabulary are the forces named). No court proceedings are reported, so court countries are unknown.

Reviewed for publication 2026-09-29: The Guardian article and the Liberty Investigates April feature were read in full from saved bodies. They come from one Liberty Investigates investigation built on an interview with Choudhury, documents shared with the Guardian and police statements, so they count as one reporting chain. The arrest date comes from the claimant’s counsel and the technical context from the National Physical Laboratory report, cited for its own contents. Claims about the arrest stay at reported status, the decisive causal question is recorded as disputed because Thames Valley Police dispute the causal account, and only the contents of the National Physical Laboratory report are recorded as documented. Choudhury is named because he spoke about the arrest under his own name to the Guardian and Liberty Investigates. The burglary suspect is not named or described beyond the reported age difference.

People described

Alvi Choudhury, a software engineer arrested at his home in Southampton (named in the Guardian and Liberty Investigates, where he spoke about the arrest)

People reported harmed in this case

1 person

0 AI participants · 1 other person harmed

One man reported arrested and held. His father, neighbours and the burglary suspect are not counted as harmed.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). Southampton: Alvi Choudhury arrested at home on suspicion of a Milton Keynes burglary after a police retrospective facial recognition match, held nearly 10 hours, claiming damages. AI incidents. https://nope.net/incidents/2026-southampton-choudhury-arrested-after-retrospective-facial-recognition-match-milton-keynes-burglary

BibTeX

@misc{2026_southampton_choudhury_arrested_after_retrospective_facial_recognition_match_milton_keynes_burglary,
  title = {Southampton: Alvi Choudhury arrested at home on suspicion of a Milton Keynes burglary after a police retrospective facial recognition match, held nearly 10 hours, claiming damages},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-southampton-choudhury-arrested-after-retrospective-facial-recognition-match-milton-keynes-burglary}
}

Related cases

Medium Unidentified facial recognition system

Aracaju (Brazil): facial recognition falsely matched a football fan to a wanted man; police led him from the stands and the state suspended the system

During the half-time interval of the Sergipe state football championship final at the Batistão stadium in Aracaju on 13 April 2024, military police took a personal trainer from the stands to a police room after the stadium's facial recognition system flagged him as a man with an open arrest warrant; he was searched and questioned and released once police confirmed he was not the person sought (G1, his lawyer and the military police). Sergipe's governor publicly regretted 'the failure of the facial recognition tool' and suspended its use until a new protocol was adopted. In June 2024 the man sued the state for 550,000 reais in moral damages; the state argued that the Aracaju municipality, not the state, had contracted the system. In December 2025 the judge rejected that preliminary objection, according to Mangue Jornalismo, which read the court file. No final judgment is reported.

Medium Metropolitan Police live facial recognition

London: Met Police live facial recognition wrongly matched a man to his watchlisted brother; officers stopped him and he later received a settlement

On 23 February 2024 Shaun Thompson, a London community volunteer whom the 2026 judgment describes as 39, walked past a Metropolitan Police live facial recognition deployment near London Bridge station. The High Court records as not in dispute that the system detected a high level of facial similarity between images of him and the watchlist image of his brother, who was the subject of an outstanding arrest warrant, and generated an alert. According to the court's account, officers stopped, detained and questioned him and made him prove his identity, and when he declined to have his fingerprints taken he was threatened with arrest; he says the experience left him angry and distressed. Officers ended the engagement after checking police records about his brother, and he confirmed his identity with a passport photograph on his phone. Big Brother Watch, whose director was his co-claimant, says he was held for almost 30 minutes despite showing several identity documents. His judicial review claim about the stop, issued in May 2024, was withdrawn under a consent order agreed on 18 December 2024 on terms including a settlement sum paid to him. An amended challenge to the Met's revised live facial recognition policy of September 2024, brought with Big Brother Watch's director, was dismissed by the Divisional Court on 21 April 2026; the court said it was not judging the merits of the technology, and the claimants announced plans to appeal.

High Unidentified facial recognition system

Maryland/Oklahoma: Kimberlee Williams, an Oklahoma woman who says she had never been to Maryland, was arrested in June 2021 and jailed about six months on bank-fraud charges in three Maryland counties after a facial-recognition search by an unknown CrimeDex respondent named her; all charges were dismissed by December 2021 and the ACLU filed complaints in April 2026

According to police and bank records reviewed by the Washington Post, in December 2019 a woman made fraudulent withdrawals at a SunTrust branch in Potomac, Maryland; branches in two other Maryland counties were also defrauded. A bank investigator posted the surveillance photos on CrimeDex, an online network of police and corporate investigators; an unknown respondent ran the image through facial-recognition software and returned Kimberlee Williams's name and photo, and the investigator told Montgomery County police that facial-recognition software had identified her. Police in Montgomery, Prince George's and Anne Arundel counties charged her in 2020 on the investigator's identification and their own visual comparison of photos (two departments also cited her old Oklahoma check-fraud convictions), without reported checks of her whereabouts; Montgomery County police did not disclose the facial-recognition search in the charging application, according to police records reviewed by the Washington Post and the ACLU. On 23 June 2021 she was arrested at the gate of Fort Sill near Lawton, Oklahoma, while accompanying her daughter on a DoorDash delivery, spent about three weeks in an Oklahoma jail, was flown to Maryland, and faced 16 charges including 12 felonies. The counties dismissed the charges one after another, in October and December 2021, after about six months in jail. Her daughter says she was depressed afterwards; the ACLU says the ordeal harmed her mental and physical health. On 14 April 2026 the ACLU and ACLU of Maryland sent complaint letters to the three departments seeking an apology and reforms.

Medium Unidentified facial recognition system

Reno casino's facial recognition wrongly matched a patron to a barred man and police jailed him for about eleven hours, his federal lawsuit alleges

Jason Killinger, a UPS truck driver, alleges in a federal civil-rights complaint filed in July 2025 that on 17 September 2023 the Peppermill Casino in Reno reported to police that its AI facial recognition software had identified him as a man barred from the casino months earlier. According to the complaint, casino security handcuffed him, and the responding Reno police officer, told the system had found a 100 percent match, did not examine his other identification and took him to jail as an unidentified person, where fingerprints confirmed who he was. The complaint says he was handcuffed for about four hours, starting in the casino security office, spent about eleven hours in police and jail custody, and was treated the next day for bruised wrists. According to the complaint, he was then charged with trespassing, and a prosecutor dismissed the charge without prejudice at the bench trial. Casino.org reports that he settled a suit against the casino. In March 2026 the court let him add the City of Reno and a municipal liability claim over facial recognition policies, but refused a new wrongful arrest claim against the officer on qualified immunity grounds. In September 2026 the ACLU, the ACLU of Nevada and the Innocence Project filed a proposed amicus brief supporting him. His September 2026 motion for partial summary judgment says a Reno police detective later concluded that the match was wrong and that he and the barred man were different people, and it cites a declaration by a former casino security officer, who worked there from about October 2024, that the system often alerted at 100 percent on the wrong people; in a declaration reported by Casino.org, the officer said he had never known the system to misidentify anyone. The case is ongoing.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents