The new European law on artificial intelligence, still awaiting final approval, will allow authorities to use facial recognition in public places.
The European Union provisionally agreed to new legislation on artificial intelligence. The specific legislation also includes a provision for live facial recognition by law enforcement authorities in public places. Amnesty International has already expressed concern, describing the move as an endorsement of "dystopian digital surveillance" across the 27 EU member states.
After intense negotiations, the European Parliament, the European Commission and EU member states have reached a tentative consensus on the Artificial Intelligence Act, which outlines the implementation of these technologies.
A key aspect of this legislation is the conditionally permitted use of facial recognition technology by judicial authorities. The European Parliament initially supported a complete ban on such real-time applications, with no exceptions.
During discussions with member states, the European Parliament compromised, allowing some exceptions. While EU officials maintain that there are strong safeguards around the use of this technology, Amnesty International, through its investigation, argues that these measures are not sufficient to prevent human rights abuses by supporting a total ban.
Exceptions provided include situations involving missing persons, victims of human trafficking or sexual exploitation, the prevention of an imminent terrorist threat and the identification of persons suspected of serious crimes such as terrorism, human trafficking, sexual crimes, murder, kidnapping, rape, armed robbery , involvement in organized crime and environmental offences.
Amnesty International's Mher Hakobyan criticizes the agreement by the three main EU institutions – the Commission, the Council and the Parliament – as a sad precedent in global AI legislation that effectively mandates invasive digital surveillance across the EU.
In addition, Amnesty International points to a major omission in the legislation: the absence of a ban on the export of harmful artificial intelligence technologies, such as those used in social rating systems.
While such systems are banned within the EU, European companies can export them.
The new bill awaits final approval by all EU member states and the European Parliament. The law will enter into force two years after approval.
