Key facts
- The report found some investigators did not know retrospective facial recognition results were intelligence, not evidence, and required further inquiries.
- Inspectors found a small number of cases where officers relied too heavily on results, leading to inappropriate investigative action.
- The report said the algorithm was more likely to wrongly flag people from ethnic minorities as suspects.
- Auditors found users uploaded poor-quality or unsuitable images, including blurred images and images with multiple faces.
- The report said legislation has not kept up with the technology and no comprehensive statutory framework governs its use.
Police officers risk arresting innocent people, especially from ethnic minorities, if they treat inaccurate facial recognition data as evidence, an official report by His Majesty's Inspectorate of Constabulary found. The report, published on Wednesday, covers only retrospective facial recognition, which is used to search images after an event rather than in real time.
The report said some investigators did not know that retrospective facial recognition results were intelligence rather than evidence and required further inquiries before officers took enforcement action. "This creates a risk that people could be incorrectly identified or arrested, which could undermine public confidence in the technology," it said. Inspectors found that in a small number of cases officers relied too heavily on the results, leading to inappropriate investigative action.
The report also found bias in the algorithm, including a greater likelihood of wrongly flagging people from ethnic minorities as suspects. Auditors identified cases where users uploaded poor-quality or unsuitable images, including blurred images and images containing multiple faces, reducing the system's ability to return reliable matches. Inspectors also found custody images that should have been deleted were kept.
"Legislation hasn't kept up with changes in facial recognition technology and there is no comprehensive statutory framework governing this technology's use," the report said. Kathryn Stone, His Majesty's Inspector of Constabulary, said police forces must show appropriate safeguards are in place, risks relating to bias and data retention are managed, and decisions remain subject to human judgment.
Jasleen Chaggar of Big Brother Watch said officers treating facial recognition as evidence rather than a probability score was alarming and that police were not logging their use of the technology. Ben Russell, lead on facial recognition for the National Police Chiefs' Council, said policing remained committed to the ethical, proportionate and responsible use of retrospective facial recognition and was pleased the inspectorate had recognised the possibilities of the technology. Prof Pete Fussey of the University of Southampton, formerly a police adviser on facial recognition, said meaningful human scrutiny of AI systems is a legal requirement, yet the report gives examples of investigators taking the computer's suggestions at face value without further checks.