This blog post describes how Alvi Choudhury was wrongfully arrested after police relied too heavily on a flawed facial recognition “match,” despite official guidance that such matches are only investigative leads. The technology is known to produce higher false positives for Asian men and younger people, and officers allegedly ignored clear red flags, including Choudhury living 100 miles from the crime scene and having no criminal record. He was detained for nearly 10 hours before being released, highlighting the risks of treating AI as decisive proof rather than using proper human judgment.
Content Authenticity Statement
100% of this week’s blog post was generated by me, the human.
The Growing Use of Police Facial Recognition Technology in the UK
I have written before on this blog about how there is neither sense nor purpose in being “Luddites” about technology in general, and specifically about the increasing use by the Police of “Facial Recognition” software. What is of overriding importance, however, is that we ensure that Artificial Intelligence is not substituted for human intelligence and common sense; AI algorithms can be useful tools, but at the end of the day they are just that – tools, not brains.
Or in other words: Fools rush in, where AI learns to tread…
College of Policing Guidance: Facial Recognition Matches Are “Investigative Leads Only”
Police Facial Recognition technology is accepted to be flawed, and the College of Policing Authorised Professional Practice guidelines make it absolutely clear that purported “matches” between faces in database A (wanted individuals) and faces in database B (individuals who for one reason or another have had a Police ‘mugshot’ taken, which is stored in the Police National Database or PND) are to be treated as ‘investigative leads only’. The algorithm’s identification of a possible match is supposed to be a starting point for critical thinking, not an end- point and not a substitute for it.
This is, not least, because of the indisputable fact that the technology demonstrates striking imbalances in the degree of “false positives” i.e mis-identifications which it produces for individuals of certain demographic groups, compared with others.
The latest overview of facial recognition technology, published by the Home Office in December 2025 has failed to assuage concerns about significant racial and gender imbalances in accuracy. In Retrospective Facial Recognition (RFR) searches results were reported which demonstrated that the “false positive” results for Asian males were 5.3% compared with 0% for White males, whilst the false positives for people aged between 21 – 29 was 5.2% compared with 0.1% for those aged 41 – 76.
APCC Warning on Built-In Bias in the Police National Database (PND) Facial Recognition Tool
Hot on the heels of that report, the Association of Police & Crime Commissioners (APCC) released a statement deploring the current state of affairs, saying, in part –
“The report from the National Physical Laboratory evaluating the accuracy and equitability of the Police National Database’s (PND) retrospective facial recognition technology tool sheds light on a concerning in-built bias. This has meant that, in some circumstances, it is more likely to incorrectly match black and Asian people than their white counterparts. The language is technical but, behind the detail, it seems clear that technology has been deployed into operational policing without adequate safeguards in place.”
The APCC statement went on to give thanks that –
“there is no evidence of adverse impact in any individual case, that is more by luck than design.”
The Wrongful Arrest of Alvi Choudhury Following a Facial Recognition “Match”
Well, as the case of my client Alvi Choudhury demonstrates that ‘luck’ has now run out – and it was Alvi who had to pay the shocking price for the same: the degradation and humiliation of a wrongful arrest for burglary.
Alvi’s case has rightly generated a lot of media interest this week. You can read his interview with the Guardian and see him describe his ordeal on Good Morning Britain.
Thames Valley Police and Hampshire Police’s Role in the Misidentification Case
I am proud to be representing Alvi and bringing my skills and expertise in this area of law to bear on the Police Forces responsible for his incarceration – both Thames Valley Police, who wrongfully requested his arrest on the basis of an alleged Facial Recognition “match”, and Hampshire Police, who carried it out. I also hope that by publicising Alvi’s story, he and I can bring pressure to bear on the Police to make them use this technology with far greater care and caution – and hopefully prevent others suffering the same fate in the future, particularly those from the most ‘at risk’ ethnic and age groups. Alvi, it will be noted, is in a cohort both in terms of his skin colour and his age which makes a false positive far more likely in his case than it would with an older, White male.
Why Policing by Consent Requires Critical Thinking, Not Blind Reliance on AI
This isn’t just a case of “racial bias” in the algorithm, however. It comes down to the problem that I highlighted at in the opening of this blog post – Policing by consent requires common sense and critical thinking, not the ‘short-cut’ of letting a computer programme do your thinking for you.
Key Red Flags Ignored in the Police Facial Recognition Identification
Sadly, I believe that is exactly what happened in Alvi’s case. The reported “match” between Alvi’s image on the PND and the CCTV still of a man wanted for burglary was taken as if it was an undisputable truth, rather than an algorithm’s ‘best guess’, and disregarded the following ‘red flags’ –
- Alvi lives over 100 miles (in Southampton) from the location of the crime (in Milton Keynes).
- Alvi is a man of exemplary character with no criminal convictions whatsoever (his image was in the Police Database only because of a previous wrongful arrest).
- Alvi’s image was taken well over 4 years before the crime – and no allowance appears to have been made for the obvious fact of almost half-a-decade of aging since then.
- The officers initially reviewing the match report appear to have only ‘seen’ the very broad racial similarities between Alvi and the wanted male (both are of Asian appearance) rather than clear disparities between them; in other words, they saw the generic, not the individual.
The Human Cost of Facial Recognition Errors and Wrongful Arrest
As Alvi told the Guardian – “I just assumed that the investigative officer saw that I was a brown person with curly hair and decided to arrest me.” Alvi was released after the Officers who came to interview him saw him – and rapidly agreed that he was clearly not the suspect in the picture; but in the meantime, he had been deprived of his liberty for almost 10 hours in Police custody.
Holding Police Accountable for Improper Use of Facial Recognition Technology
To my mind, Alvi’s arrest was as a result of the Police playing “AI lottery” with people’s lives, and I intend to make them pay the price for that – and in doing so, help to drag improper uses of this technology into the full light of day.
How you can help
This blog exists to inform, empower, and reassure people who may otherwise feel overwhelmed when dealing with unlawful police conduct. If you’ve found this article helpful or informative, I would be very grateful if you could leave a 5-star review. Your support helps me continue this work and, just as importantly, helps others find reliable, specialist advice when they need it most. Thank you.
Contact the Expert Police Misconduct Solicitor
Iain Gould is a solicitor specialising in complaints, claims and civil actions against the Police. With over 30 years of experience and a national reputation, he has successfully sued all 43 police forces in England and Wales challenging abuse of power and securing rightful compensation.
You must be logged in to post a comment.