Police facial recognition is now highly accurate, but public awareness lags
2 days ago
- #policing-reform
- #facial-recognition
- #public-trust
- UK government plans to increase facial recognition technology use in policing, expanding live facial recognition vans from 10 to 50.
- £26 million allocated for a national facial recognition system and £11.6 million for live facial recognition technology.
- Facial recognition has led to 1,700 arrests in the Met police, according to Home Secretary Shabana Mahmood.
- Public awareness of facial recognition technology's accuracy and use remains low, with only 10% feeling confident about their knowledge.
- Facial recognition is used in three main ways by UK police: retrospective analysis, live public space scanning, and operator-initiated checks via mobile apps.
- Accuracy of facial recognition has improved, with false negative rates below 1% and false positives at 0.3%, as per US National Institute of Standards and Technology.
- Historical biases in facial recognition, such as higher error rates for non-white faces, have been reduced with updated, balanced training datasets.
- Public support for police use of facial recognition remains high for criminal investigations (89%) and locating missing persons (89%).
- Concerns exist about public trust, with only 55% trusting police to use facial recognition responsibly, down from 63% in 2020.
- Calls for a legal framework covering all facial recognition users to maintain public trust and ensure responsible use.