top of page

Reversed Surveillance

In January 2023, the French government passed a bill allowing the use of algorithmic video surveillance (AVS) as a legal crowd controlling measure for any gathering of more than 300 people. This technology is put in place to automate the process of detecting crime, as it is able to capture features that are missable to the human eye. Artificial intelligence decides who is a suspect and who isn’t.


Emotion recognition is similar to facial recognition, but instead of focusing on identity recognition, its aim is to decode the person’s inner emotional state. In this case, the artificial intelligence is trained to detect, analyse and decode human behavioural patterns associated with emotions.


Starting from a 3 hour long livestream of a recent French protest, the artist uses the software ‘Video Content Analysis’ to analyse the crowd and discern over 30 000 people’s faces. 

With the use of an AI trained to recognise 7 facial expression associated with determined emotions the artist categorises the protestors based on the emotional state they were assigned by the artificial intelligence. This process allows him to visualise how the AI interprets a protest.


There is currently no record of emotion recognition being used in decisions affecting protestors. However, this technology has been widely used by companies to monitor their employees.


The continuously evolving nature of surveillance technologies, combined with the normalisation derived by the ease with which these technologies are introduced on a legislative level, poses a threat to the right of assembly, protest, and speech. This becomes extremely threatening when the technology is given the task to decode something of human nature, and used to make decision based on probability rather than events.

bottom of page