One step closer – the EU AI Act moves to final stages with a ban on facial recognition in public places

Why should I read this?

Here we set out the details of the latest update on the EU AI Act, which has moved another step closer to becoming law with a vote passed on Wednesday 14th June to move the draft text to the final stages of individual negotiations with members of the European Parliament (MEPs).

Ahead of the vote, the Act’s ban of the use of facial recognition technology in public places faced a last minute challenge from the Christian Democrats of the European People’s Party (EPP) to allow certain exemptions to this ban in the case of law enforcement, for example identifying suspects of crimes or investigating terror threats.

As a reminder – the draft AI law says some AI is banned; some is high risk; some does not pose high risk. The last minute challenge on the use of facial recognition technology picks up on the opposing views across political parties and activists with regard to the use of technology by governments. Groups such as Amnesty International have raised concerns over the use of invasive facial recognition technology that “amplifies racist and discriminatory law enforcement against racialized people, including stop-and-search practices which disproportionately affect Black and brown people. It is also used to prevent and curtail the movement of migrants and asylum seekers” – a concern that was echoed by some MEPs.

Romanian MEP Dragos Tudorache echoed the anti-facial recognition sentiment: “We don’t want mass surveillance, we don’t want social scoring, we don’t want predictive policing in the European Union, full stop. That’s what China does, not us”.

What does this mean for regulation of AI?

With the EU AI Act being seen as a benchmark law which others may look to follow, this tension and debate shows just how difficult it can be to decide between banning certain uses of AI to protect against a particular type of harm, whilst at the same time protecting the public from other serious harm and encouraging innovation.

MEPs ultimately voted in favour of banning the use of facial recognition technology in public places. Brando Benifei, a member of parliament from Italy, said after the vote:  “While Big Tech companies are sounding the alarm over their own creations, Europe has gone ahead and proposed a concrete response to the risks AI is starting to pose”.

It appears that the EU may be erring on the side of caution when it comes to AI, no doubt influenced by the rise of disruptive technologies such as generative AI, which caused more than 1,000 tech leaders, including the likes of Elon Musk and Apple co-founder Steve Wozniak, to call for a six-month pause on the development of the most powerful artificial intelligence systems.

This EU AI Act has become a symbol of the fine line between regulation and innovation, the latter being one of the purposes of the Act to begin with – not to stifle innovation, but encourage it. Critics might say the vote to ban the use of facial recognition in public  places, without considering its benefits in the criminal and terrorist threat space, is too restrictive a stance.

These points of contention are perhaps one reason why the UK seems to be adopting a wait-and-see approach with its consultation on the white paper (a principles based approach rather than legislation – for now).  

Might the EU stifle innovation by choosing a more stringent approach on AI? Is it right to go that way even if that AI might help prevent crime and terrorism? And can it with this benchmark new law be confident that it has balanced protection from harm which is caused by AI, against the safety concerns of society more broadly?  A lot of questions.  No doubt lots of differing opinions. Which way might other countries and regions go on this particular safety concern? There’s a lot to be figured out as we all move into this new space of regulatory AI compliance.  Is it fair to say that, with the rise of generative AI technologies, the EU must lead the way in not allowing AI to run away with itself. 

Now that the draft text has been agreed, the Act will move to a three-way negotiation  process (known as a trilogue) between the European Parliament, EU Council and European Commission. The first trilogue session took place on 14th June, the same day the draft text was agreed.

For more information about the EU AI Act

Stay with us for further updates on the progression of the Act and AI regulation discussions. We have our hotly anticipated EU AI podcast series coming soon…

Further reading on AI

• Upcoming vote on EU AI Act
• European Commission approves next version of AI Act
• The march to regulatory change for artificial intelligence: the commonalities between the EU and US
• UK government’s white paper on AI – innovation balanced with protection from harm
• AI UK white paper and regulatory guidance – the ICO response
• Eversheds Sutherland webinar “Demystifying generative AI and the legal issues it raises” – AM session and PM session.