France’s parliament has approved a plan to use artificial intelligence to monitor public spaces during the 2024 Paris Olympics. The system is intended to identify suspicious behavior, but critics say that goal is too vague and could bring extensive surveillance of people attending the games.
What France approved
The vote took place on Thursday, following an earlier approval by the French Senate. Around 40 MPs had signed an open letter opposing the proposal. The 2024 Olympics Games are due to take place in Paris between July 26 and August 11.
The measure would use automated behavioral surveillance, applying AI to activity in public places. The source describes the system as looking for “abnormal behavior.” That broad standard is central to the criticism: assessing behavior is different from identifying a specific person or responding to a clearly defined threat.
How the plan intersects with EU AI rules
The EU’s AI Act is a proposed, risk-based framework for regulating AI applications. Its original draft includes a prohibition on real-time remote biometric identification in publicly accessible spaces for law enforcement, with exceptions for certain searches and serious threats. Members of the European Parliament have pushed for a more comprehensive ban.
That framework was still under negotiation when the French parliament voted. Its final details were unsettled, and even a speedy compromise might not have taken effect in time for the Olympics. As a result, the French plan sits amid uncertainty about how the eventual EU rules will apply.
Critics argue that the proposal reaches beyond the narrow exceptions contemplated in the draft AI Act. Their concern is that a system designed to flag undefined suspicious behavior could monitor people broadly, rather than being limited to specific cases.
Critics warn of errors and discrimination
Patrick Breyer, an MEP in the European Parliament with the Pirate Party, called the technology “error-prone” and intrusive. He argued that automated systems could wrongly flag many people, discriminate, and encourage conformity in public spaces.
Those concerns point to a practical problem with behavior-monitoring systems: a flag is not proof of wrongdoing. If the criteria for “abnormal” activity are unclear, people may not know why they were identified or how to challenge an assessment. The source reports Breyer’s view that such systems are ineffective at catching criminals, while the plan’s supporters’ case is not detailed in the article.
The dispute therefore concerns more than technical accuracy. It also asks how much observation people should expect in shared spaces, and what safeguards should apply when an algorithm’s assessment can draw attention to an individual.
Legal questions remain
The French plan could still face a challenge before the country’s constitutional court. The source also notes that the CNIL, France’s data protection watchdog, had recently set up a department focused on artificial intelligence in preparation for the EU AI Act. Its view of the Olympics plan was not yet reported.
The vote also reflects a broader tension described in the source: national security priorities can conflict with EU protections for fundamental rights. France has previously resisted EU rules on general and indiscriminate data retention, arguing that such activity is essential for national security, while the bloc’s top court has found fault with bulk data collection regimes.
Whether Olympic attendees will ultimately be assessed by algorithms remained uncertain. The unresolved constitutional challenge, the evolving EU framework and scrutiny from data protection authorities could all shape what happens next. The decision has put a larger policy question in focus: how governments can use AI in public spaces while limiting intrusive or indiscriminate monitoring.