Visual Privacy: Current and Emerging Regulations Around Unconsented Video Analytics in Retail
It addresses privacy and legal challenges for retailers using video analytics, highlighting an incremental analysis of evolving regulations.
This research examines the current and emerging legal regulations concerning unconsented video analytics in retail, where machine learning models can infer personal characteristics like age and health from video data, and advises retailers to adopt a cautious approach to avoid legal and brand risks.
Video analytics is the practice of combining digital video data with machine learning models to infer various characteristics from that video. This capability has been used for years to detect objects, movement, and the number of customers in physical retail stores, but more complex machine learning models combined with more powerful computing power has unlocked new levels of possibility. Researchers claim it is now possible to infer a whole host of characteristics about an individual using video analytics, such as specific age, ethnicity, health status and emotional state. Moreover, an individuals visual identity can be augmented with information from other data providers to build out a detailed profile, all with the individual unknowingly contributing their physical presence in front of a retail store camera. Some retailers have begun to experiment with this new technology as a way to better know their customers. However, those same early adopters are caught in an evolving legal landscape around privacy and data ownership. This research looks into the current legal landscape and legislation currently in progress around the use of video analytics, specifically in the retail store setting. Because the ethical and legal norms around individualized video analytics are still heavily in flux, retailers are urged to adopt a wait and see approach or potentially incur costly legal expenses and risk damage to their brand.