France approves algorithmic video surveillance to safeguard Olympics

France has legalised the use of intelligent video surveillance for major events including the 2024 Paris Olympics and Paralympics – despite warnings the technology is a blatant breach of privacy that violates civil liberties.

Privacy fears have been raised over a law allowing police and security companies to monitor crowds at large-scale events, including the 2024 Paris Olympics, using video surveillance powered by artificial intelligence. AFP - THOMAS SAMSON

Following a vote late Tuesday in the National Assembly, a law to trial video surveillance powered by artificial intelligence (AI) was adopted by a large majority. It takes effect immediately and will remain in place until the end of 2024.

Algorithm-driven “smart” cameras capable of detecting crowd surges, fires, abandoned bags, and unusual or risky behaviour are to help ensure the safety of millions of spectators travelling to Paris for the Games.

The government argues the cameras, some of which will be mounted on drones and aircraft, will be able to spot potential dangers in real time, allowing such risks to then be communicated to police or other security services.

Biometric or not?

While those behind the law insist facial recognition is off the table and no biometric data will be processed – a condition laid down by France’s CNIL privacy watchdog – opponents say any data that identifies a person by their physical attributes is by definition biometric.

“This is a massive surveillance tool that allows police to analyse our behaviour … and to decide who is normal and who is suspicious based on their own stereotypes,” Noémie Levain, a legal adviser at the digital rights group La Quadrature du Net, told RFI.

“It can detect what the police want it to detect.”

This includes the way a person walks, the colour of their skin and their gender, Levain says, adding that the technology itself is made by private companies and is shrouded in secrecy.

Beyond the Olympics

The so-called “Olympics security law” will also apply to large-scale sporting, recreational and cultural events happening both before and after next year’s Games – including the Rugby World Cup which kicks off in September.

Any event with more than 300 participants will in theory fall under the scope of algorithmic monitoring.

“This is going to be tested on concerts, festivals, races and the Rugby World Cup – so French people are going to be the guinea pigs to teach the algorithm,” says Levain.

• Covid rules rebooted as France embraces QR codes
• France fines Google, Facebook for making cookies difficult to refuse

Leftwing opposition MPs in France unsuccessfully pushed to have the AI surveillance restricted to objects and not people, and to have the resulting data analysed by the state and not by private companies.

Both the right and the far right threw their support behind the bill.

In a report to lawmakers, La Quadrature du Net argued that sensitive biometric information collected by the cameras would infringe on the European Union’s own General Data Protection Regulation – the toughest privacy and security law in the world.

EU-wide debate

French policy is being closely watched by MEPs in Brussels who have spent the past two years debating the AI Act – legislation governing the use of artificial intelligence by states and private companies in Europe.

“France has been pushing for the least protections … which is very worrying,” says Levain from La Quadrature du Net, adding that many companies pioneering AI technology are French and stand to profit from changes to the law.

“A lot of European members of parliament are willing to ban massive biometric surveillance, so we hope that European values and European vision of liberties will win and that France will realise it is an isolated country.”