The European Union’s Artificial Intelligence Act has reached its final stages, with EU countries reaching an agreement on its technical details last Friday. The provisional agreement on this landmark legislation now awaits approval from European Parliament committees. However, despite expectations of significant support, some legal experts are already voicing concerns about gaps in regulating remote biometric identification, which could potentially infringe upon human rights.
Douwe Korff, Professor of International Law at London Metropolitan University, highlighted potential issues within the final text of the Act from a human rights perspective. One notable concern is the treatment of remote biometric identification systems, which can be developed and tested in real-world conditions without the requirement for an impact assessment on fundamental rights or registration in a relevant database. This loophole could potentially permit the use of such systems in emergency situations, Korff warns.
Korff also criticized the AI Act’s requirement for notification to relevant authorities when real-time biometric identification is utilized. He noted that the Act does not specify that this notification must occur simultaneously with the request for authorization, potentially allowing law enforcement agencies to notify authorities after the systems have already been employed. This lack of real-time oversight, according to Korff, may undermine the effectiveness of supervision.
The next crucial step for the AI Act involves votes from the internal market and civil liberties committees, expected on February 13. Formal approval by the European Parliament is anticipated on April 10 or 11, after which the legislation would come into force 20 days following its publication in the official journal, EuroNews reports.
Despite the progress, concerns persist regarding potential amendments from privacy-focused lawmakers, which could potentially delay the legislative process. Moreover, uncertainty surrounds how the Act will be implemented at the national level, with reports suggesting reservations from certain EU member states, including France, Austria, Germany, and Italy.
Following the agreement, the European Commission released a statement reminding member states of their ability to adopt more stringent rules for technologies like facial recognition, emotion recognition, and biometric categorization. Additionally, plans were announced to establish an expert group comprising authorities from EU member countries to advise and assist the Commission in implementing the AI Act.
The full text of the provisional agreement has been published by the Council of the EU, marking a significant milestone in the EU’s efforts to regulate artificial intelligence.