BFMTV report sparks online reaction
A discussion on artificial intelligence and predictive policing during a BFMTV report on October 20 prompted the X AlertesInfos account to post a sensationalist tweet claiming France was experimenting with AI software to anticipate riots. The tweet, which gained significant attention, referenced the 2002 science fiction film Minority Report, which depicted a system capable of predicting crimes.Mock demonstration created online
In response, internet users created a mock demonstration in Clermont-Ferrand, which was not actually planned. The event highlighted the public’s reaction to the idea of AI being used to predict and prevent unrest, even though no such plan existed.Predictive policing tools and their limitations
The use of predictive policing technologies, such as the American Palantir and the French ChapsVision, has raised questions about their ability to predict spontaneous gatherings and prevent violence. A researcher, who wished to remain anonymous, noted that while these tools can identify trends, human intelligence remains more effective for now.How predictive policing works
Predictive policing relies on software that analyzes large amounts of data, including surveillance camera images, crime histories, criminal records, and social network data. After the November 13, 2015 attacks, the French government used Palantir before switching to ChapsVision's ArgonOS software in June 2026 to anticipate terrorist threats.Transparency and concerns
However, the functioning of these tools remains opaque, as companies and state services are often reluctant to provide information to journalists and researchers. A report by La Quadrature du Net highlighted the lack of transparency surrounding the "Smart Police" software, noting that it is primarily used for digital management and coordination of municipal police rather than for detecting future crimes.Global use and mixed results
PredPol, an American predictive policing program, was used by the Los Angeles police department from 2010 until it was discontinued in 2020 due to budgetary issues and concerns about racial discrimination. An investigation found that less than 1% of the predictions made by PredPol in New Jersey in 2018 were accurate.Argentina’s AI security unit
In August 2024, Argentina established the Unit of Artificial Intelligence Applied to Security (UIAAS) to monitor social networks, websites, applications, and the dark web. Despite the global interest in predictive policing, no one has been able to quantify the effectiveness of such tools. The symbolic dimension of these technologies is considered significant, as they may influence political agendas and the development of new laws.
