Gian Volpicelli of Politico reported last week that the AI Act had undergone last-minute changes, allowing law enforcement to use facial recognition technology on recorded video without judicial approval.
The original agreement had foreseen the use of post-facial recognition only for very serious crimes, under stringent conditions. The latest version of the text (helpfully leaked by Luca Bertuzzi of Euractiv), however, would allow law enforcement agencies to employ non real-time facial recognition following approval from an administrative authority, rather than a judge's decision. Critics have also expressed concern that the technology would be authorized to identify suspects for all types of crimes, irrespective of the severity of these offenses.
The alert was sounded by German MEP Svenja Hahn who criticized these modifications in the final text (which was agreed upon in December), describing them as an assault on civil rights and drawing parallels between the potential misuse of biometric technology and practices in China. Hahn contends that the changes deviate gravely from the original agreement, which mandated stricter conditions and judicial oversight for facial recognition use.
Hahn is supported in her pushback by civil and digital rights associations such as EDRi and https://reclaimyourface.eu/