Germany positioned to dig heels in over EU’s AI Act, biometrics use
Germany is positioning itself more closely with the European Parliament rapporteurs’ aims for stricter protections in areas such as biometric recognition and predictive policing within the upcoming AI Act, falling out of line with other European Union members ahead of discussions between the bodies of the bloc, according to analysis by Euractiv.
Although even Germany’s stance is more nuanced according to submissions seen by the outlet.
EU alignment to ‘open door to surveillance’, civil society aligned in push back
At a meeting on 6 December 2022, EU Council ministers compromised and agreed to a common position on biometrics in the AI Act, including elements such as allowing remote biometric identification in public places by law enforcement when strictly necessary.
Civil society organizations criticized the move and a perceived weakening in the classification of what are “high-risk” AI systems used in surveillance. One hundred and sixty-four organizations and 30 individuals signed a joint letter to demand better protections from AI and biometric surveillance, particularly for vulnerable groups such as migrants at borders.
“The position adopted today would enable a dystopian future of biometric mass surveillance in Europe, potentially exposing everybody to constant identification, monitoring their behaviour and analysing their emotions in public spaces,” wrote German Member of the European Parliament Patrick Breyer (Pirate Party, Greens/EFA group) at the time.
“It would justify the permanent and ubiquitous deployment of face surveillance to look for the thousands of ‘victims,’ ‘threats’ and suspects of ‘serious crime’ that are wanted at any time.”
German position going into the trilogues
The European Parliament seems on the side of supporting a ban on biometric mass surveillance both in public and private places, in real-time and retrospectively. This would also cover emotion recognition, lie detection and crowd control.
Back in December, the German minister at the meeting noted that “there is still room for improvement,” reports Euractiv, and that he hoped his country’s points would be taken into account at trilogues on the AI Act this year.
The trilogues are informal negotiations between the European Parliament, the Council of the European Union and the European Commission. They would aim for the Parliament and Council to reach provisional agreement, which each of the co-legislators would adopt.
Along with the Parliament’s rapporteurs, Germany had been calling for a total ban on biometric recognition technology. But Euractiv has seen documents submitted by Berlin that show it is only in favour of banning real-time biometric identification in public, allowing retrospective recognition.
Germany also wants to shift to the GDPR’s definition of biometric data to prevent a divergence of terminology and risk level classifications.
Berlin wants a ban on any AI application that substitutes human judges in assessing the risk of an individual committing or repeating a crime. The Council added these as high-risk categories, Parliament wants a ban. Germany wants a ban on authorities using AI tools such as polygraphs or emotion recognition.
Germany has also lobbied to prohibit AI systems monitoring employee performance. The country’s proposals may not have made it to the Council-agreed text, but German representatives could be supportive of Parliament in negotiations.