Swedish proposal tests AI Act’s live public facial recognition limits

In a move that has sparked debate and raised concerns about privacy and civil liberties, the Swedish police are considering the use of cameras equipped with real-time facial recognition to identify suspected criminals in public spaces. The proposed system would automatically scan the faces of individuals against a database of known offenders and alert law enforcement if a match is found.

Proponents of the technology argue that the use of facial recognition cameras could help law enforcement agencies quickly identify and apprehend dangerous criminals, thereby enhancing public safety.

The proposal suggests that the police, in accordance with the EU’s AI Act, should be allowed to use camera surveillance with automatic facial recognition in real time in public places, NyTeknik reports.

The proposal was tabled in December, and The Local reports that Sweden’s government then called an inquiry, headed by Kazimir Åberg, to determine the legality of the proposal under the AI Act. The Act bans real-time remote biometric identification in public spaces, with exceptions handled at the member state level, and noted in an analysis of the AI Act by TechLaw Sweden.

Åberg found that live facial recognition could be activated in the country, but with a watchlist populated only by people accused of crimes with prison terms of four years or longer, and with judicial authorization.

If approved, the technology will also be allowed to search for missing persons.

A separate proposal means that the police would have access to footage from road traffic cameras to investigate terrorist crimes, as well as a series of other crimes that can result in at least three years in prison.

The proposals come in the wake of gang violence in the country, as reported last year, pushing the government to consider surveillance changes. The country’s data privacy watchdog released an opinion on the police using biometric data stored in the Swedish passport register to crackdown on crime.

However, critics of the proposal raise ethical and legal concerns, arguing that the use of facial recognition technology poses a threat to individual privacy. In 2021, the Swedish authority for privacy protection determined that the Swedish police authority violated the Swedish Criminal Data Act by using Clearview AI for identifying individuals.

Article Topics

AI Act  |  biometrics  |  criminal ID  |  facial recognition  |  police  |  real-time biometrics  |  Sweden