According to NBC News, a Japanese startup called Vaak is using artificial intelligence software to help retailers ferret out shoplifters before they leave the store … and perhaps even before they grab the goods.
The “Vaakeye” technology is designed to work with a standard surveillance system to help identify behaviors that may constitute “suspicious activity.” Vaak CEO Ryo Tanaka told NBC that this could include facial expressions, gestures, and movements — even clothing choices. If someone sets off an internal alert, the idea is for the store to dispatch personnel for a quick check-in.
And while Vaak says its tech reduced shoplifting incidents by 77% in tests, the idea of using technology to solve the problem is not without controversy. Sven Dietrich, a professor at John Jay College of Criminal Justice, told NBC that these kinds of deep learning algorithms are only as good as the data used to train them and “might be extracting a certain bias.”
Likewise, a recent report by CNet suggests that systems that use facial recognition to ID shoplifters can be inherently problematic, citing California-based video surveillance startup Kogniz and its “multi-location” cloud system that can save the image of a shoplifter and share it with other stores’ locations.
According to the report, customers can then be flagged whether or not a crime has been committed, and could even find themselves barred from transacting with private businesses. Says CNet, “One mistake could mean never being able to shop again.”
Because there are few standards at this point regarding how facial recognition can and should be used, the piece sounds alarm bells over what could be seen as overreach: “It’s just one example of how facial recognition straddles the line between being a force for good and being a possible violation of personal privacy.”