During the Olympic fortnight, 485 surveillance cameras were controlled using artificial intelligence software, as an experiment. Algorithmic video surveillance (AVS) is a technology that works with the existing network of video surveillance cameras. The goal is to make them “intelligent” passing the captured images through software that uses artificial intelligence. Currently, the objective is not to recognize faces, but to detect situations considered as “suspicious”. They could be crowd movements, falls, fire, weapons, etc.
Until now, this image analysis work was mainly carried out by video surveillance operators, who then alerted the police. The promise of this technology is to help these agents process an exponential amount of video surveillance images.
For its detractors, this artificial intelligence-driven surveillance raises numerous questions, particularly in terms of respect for privacy and discriminatory biases.
In this video, we return to how algorithmic video surveillance works and its challenges. We also interviewed Katia Roux, advocacy officer at Amnesty International France, responsible for the questions “technologies and human rights”on the issues raised by this new technology. To learn more, we invite you to read the article below on biometric recognition in public spaces.
“Understand in three minutes”
The explanatory videos that make up the “Understanding in three minutes” series are produced by the Vertical Videos department of the World. Mainly spread on platforms such as TikTok, Snapchat, Instagram and Facebook, their goal is to contextualize big events in a short format and make the news accessible to everyone.
