While the National Commission for Computing and Liberties (CNIL) frequently deals with subjects related to video protection and video surveillance, the personal data protection authority now wants to draw attention to the potential dangers of the deployment of so-called “intelligent” or “augmented” video devices in the public space.
Used in particular in industry, defence, connected cities and retail, smart cameras based on artificial intelligence are seeing their market grow rapidly. However, their growth is not without risks.
To put it simply, it is a video protection device coupled with automated image processing software, with the aim of analyzing and interpreting what is captured in real time without resorting to the human eye. .
The software overlay makes it possible to recognize objects, silhouettes or movements in a probabilistic way. Such a device can in particular make it possible to calculate the frequentation of a metro train, to obtain statistics on compliance with the wearing of a mask or even to detect crowd movements or abandoned objects.
Augmented device does not necessarily mean facial recognition
If, for the CNIL, these cameras raise “new issues for the rights and freedoms of individuals”, they must however be distinguished from biometric camera systems.
Indeed, unlike cameras based on facial recognition, augmented cameras do not necessarily process the characteristics (physical or behavioral) of individuals, nor are they used to identify the people filmed. However, these two criteria are always found in biometric recognition devices, summarizes the CNIL.
Simply put, the purpose of an augmented camera is to categorize and analyze through AI without uniquely identifying a person. A biometric camera aims to identify or authenticate a person by comparing the images captured in real time with a bank of pre-existing images.
The CNIL warns about the risk of surveillance
And this distinction has real legal consequences. Thus, the CNIL emphasizes that biometric recognition devices involve “sensitive” data processing which is in principle prohibited by the GDPR for the vast majority of them.
With regard to augmented video devices, on the other hand, the CNIL recognizes that the risks are more “graded”, depending on the resulting use. The risk to individuals will be higher if a camera detects violations, for example, or enables targeted advertising with passers-by. On the contrary, if a smart device aims to make statistics on the use of a metro, the level of intrusion into the private sphere will be lower.
The CNIL warns of the poorly controlled generalization of these technologies, which are by nature “intrusive”, and not yet subject to a specific legal framework.
According to the regulator, the red line would be to use these smart cameras for the purpose of “scoring” people. The CNIL therefore calls for “an overall reflection on the fair use of these tools in the public space”.