With the Artificial Intelligence Act (AI Act), the European Union aims not only to regulate artificial intelligence but also to be a global standard-setter, writes Lorenzo Ancona. The proposed text of the regulation encompasses different artificial intelligence systems, such as generative intelligence (e.g. ChatGPT and Midjourney) and biometric intelligence.

For the past two years, The Good Lobby, together with the ‘Reclaim Your Face‘ movement, has been closely monitoring the negotiations of the European institutions and have advocated for a comprehensive ban on artificial intelligence for mass biometric surveillance.

Following the tension between the Council of the European Union and the European Parliament between the end of 2022 and the beginning of 2023, biometric surveillance is addressed through a compromise solution: the current proposal envisages a complete ban to real-time mass biometric surveillance while limiting ex-post surveillance only to serious crimes and with pre-judicial approval. On 11 May, the Civil Liberties Committee (LIBE) and the Internal Market Committee (IMCO) voted for a complete ban on artificial intelligence for biometric recognition, including real-time and ex-post uses. Now, the text will be voted in the assembly and then discussed within trilogue.

In this article, we will summarise the negotiations and proposals of the last months that led to the compromise solution.

Raising awareness of invasive practices

Although the first text proposed by the European Commission in April 2021 contained a ban on mass biometric surveillance, several organisations – such as Reclaim Your Face – had demonstrated dissatisfaction with the removal of the prohibition of the ex-post use of collected data and other limitations. Our activities draw attention to the issue and resulted in the commitment of 24 MEPs, who proposed amendments for a total ban on biometric identification in public spaces in June 2022.

In October 2022, the MEPs Brando Benifei and Dragoș Tudorache, the co-rapporteurs on the AI Act text, proposed to rephrase some articles on prohibited practices, such as social scoring and mass biometric surveillance. In stark contrast to the original text of the AI Act, Benifei and Tudorache aimed to remove all the exceptions from the Commission text, extending the ban to ex-post biometric recognition, in addition to real-time recognition. Also, they sought to include public spaces in the provision, in addition to private and online spaces. This proposal also provoked friction at the political level, with the EPP especially opposed to the proposed total ban.

Subsequently, in December 2022, the Czech-led EU Council took a stance opposite to the co-rapporteurs in Parliament. The Council not only ruled out a ban on ex-post biometric recognition but also proposed increasing the exceptions for the real-time use of this technology, particularly when needed by law enforcement agencies for security reasons. This decision was widely criticised by Reclaim Your Face, which called it ‘disappointing’ and a first step towards a legal basis of protection for invasive practices, as well as against data protection laws and human rights in the EU.

In favour of a comprehensive ban

After the Council’s statement, at the beginning of 2023, the discussion on biometric surveillance resurfaced in the European Parliament for the assembly vote on the amended AI Act text. Ultimately, on 11 May, the Civil Liberties Committee (LIBE) and the Internal Market Committee (IMCO) voted in favour of a comprehensive ban on artificial intelligence for biometric recognition, encompassing both real-time and ex-post applications. The ban will effectively prohibit mass facial recognition programs in public spaces and predictive policing algorithms utilising personal data to identify potential future offenders, with exceptions granted only for ex-post technologies employed in the prosecution of serious crimes and under judicial authorisation. Furthermore, the approved regulation requires the creation of a public database including “high-risk” AI systems deployed by public and government authorities to inform citizens properly.

A discussion among MEPs is now scheduled for 13 June, followed by a vote in plenary on 14 June. Then, trilogue will commence.

We will keep acting, along with Reclaim Your Face, to ensure the implementation of the complete ban, as the risks associated with using biometric data, in terms of security and privacy for citizens, are far too significant to refrain from a relentless advocacy effort.

(The content of this post is based on a previous blog post by Lorenzo Ancona, which was published in Italian)