Despite a recent controversy, UK Police Forces are planning to continue with the rollout of Facial Recognition Technology deployments across the country. While this technology undoubtedly makes it easier and quicker for the police to identify potential suspects, critics and civil liberties groups remain unconvinced of the need for mass surveillance in a democratic state, and the potential infringement of civil rights. What are some of the challenges facing widespread deployment of facial recognition by the police?
What types of facial recognition systems do the police use currently?
According to a government factsheet, police forces in the UK currently use three main types of Facial Recognition Technology:
Operator Initiated Facial Recognition (OIFR):
A police officer can take a photo of a person of interest and check their identity without the need to arrest them. This is useful when they are not certain who the person is.
Retrospective Facial Recognition (RFR):
This is used to review video footage after an incident. The footage and images can come from CCTV, dashcams, video doorbells, and mobile phone footage. The faces in these videos are compared against arrest photos of a suspect to identify them.
Live Facial Recognition (LFR):
Live video footage is captured when people are near an active facial recognition system. This is in a public place, and the police make the public aware that Live Facial Recognition technology has been deployed. It compares the live footage against a pre-prepared watchlist.
Tensor's facial recognition systems can be used as Live Facial Recognition or as Retrospective Facial Recognition, as well as Live Facial Recognition implementation for some of our clock-in scanners and access control terminals.
Why is public opinion important?
Policing in the United Kingdom is based on the 'policing by consent' concept. This means that the police have the powers required to uphold the law, but only with the consent and approval of members of the public, and not because government has given them the powers. This means that the police must be able to justify their actions and be held accountable if they cannot.
Because of this, the police must consider public reactions and opinion towards any new policies or technologies they want to implement. With Facial Recognition Technology, recent studies were cited that show there was positive public support. According to a government publication discussing public attitudes to the police using Facial Recognition Technology, there are two previous surveys referenced which show this. In 2019, 70% of the UK population was supportive of the police using Facial Recognition Technology for criminal investigations, and in 2025, 91% of the UK population thought the use of Facial Recognition Technology was beneficial to policing.
What was the recent controversy?
Essex Police had been using a Live Facial Recognition system since the summer of 2024 after a successful trial in 2023 but, in March 2026, was forced to pause its usage to investigate an issue of potential identification bias.
The identification bias report said the system was more likely to correctly identify men than women, and "it was statistically significantly more likely to correctly identify black participants than participants from other ethnic groups".
Unfortunately, no system is 100% infallible. However, there are many measures that can be put into place that can help mitigate risk or bias. Biases are sometimes harder to identify and overcome, especially if they are internalised or unconsciously influencing actions. It is important to remove biases and risks to make sure that everyone is treated fairly.
In this example, risk and bias mitigation relies on testing the recognition algorithm used to identify facial biometric markers and checking the rates of different metrics. Essex Police had stated that the False Match Rate (FMR) needed to be at 0.6 or above to achieve a fair rate of false positive identification. However, they had potentially repeated a claim that did not apply to the Facial Recognition Technology system they had selected. The Live Facial Recognition they had used was not the same as the one used by other UK police forces. Essex Police chose to pause the deployments of their Facial Recognition Technology system and review the information with the company that developed the software. This led to an update to the software, and improved policies and procedures during use.
Will this impact future police Facial Recognition Technology deployments?
In January 2026, Home Secretary Shabana Mahmood announced plans to increase the total number of Live Facial Recognition vans available to UK police forces from ten to fifty. These plans are still going ahead as part of the government's strategy to "build a safer society for women and girls", and public support remains high.
Most of the criticism for police Facial Recognition Technology use centres on the right to an individual's privacy, and how biometric data is kept or stored afterwards. One solution is to delete the data after a negative Facial Recognition Technology match.
When handling any kind of personal data, guidelines and policies must be followed to make sure that it is only used when required and stored securely or disposed of afterwards. Facial Recognition Technology and other biometric data need to be treated with care and in accordance with personal data protections laws and standards, like GDPR. Read more about this in our blog about UK GDPR and biometric data.
With Tensor's Facial Recognition CCTV and Facial Recognition Access Control systems, we can record evidence quality footage, with clear and secure audit trails. This means there is a full trail of responsibility and accountability for the data that is used and kept.
Talk to our experts to find out how facial recognition can benefit your business.