Using artificial intelligence to curb shoplifting< return
Every day, we find ourselves fascinated with another revolutionary development in artificial intelligence. It may be an artistic creation, a medical breakthrough or a type of technology that we will soon install in our homes. Police services are increasingly putting artificial intelligence to good use.
A number of police forces and agencies are already using it to solve crimes. For example, some 200 U.S. agencies are using an algorithm from the University of Southern California that combs through 25 million web pages to find victims of human trafficking or sexual exploitation. The London police also want to use AI to predict violent crime before it even occurs.
Along the same lines, technology and artificial intelligence are now being used to tackle shoplifting. In 2017, the retail industry lost $34 billion due to theft. That amounts to a 2% loss in revenue, which is huge for a store that is already struggling with a number of issues.
A Japanese company called Vaak has shown that its technology can identify shoplifters even before they steal something. The company’s algorithm analyzed 100,000 hours of surveillance camera footage of shoplifting incidents. The system, called VaakEye, has compiled a wide range of data, such as facial expressions and clothing. It alerts employees via a mobile app. Salespeople can then ask the person if they need help or are looking for a particular product. This approach would prevent many would-be thieves from following through with their plans.
Vaak’s technology was tested in 50 shops in Japan. According to the company, the shoplifting rate dropped 77% during the testing phase. The facial recognition system and artificial intelligence developed by Vaak could ultimately be used to better understand consumers’ habits in order to help retailers become more efficient. In public places, Vaak could also identify suspicious behaviour in people and even prevent suicides.
Many questions still need to be answered
Vaak hopes to see its technology implemented at 100,000 stores in the next three years. This innovation is sure to please retailers. Consumers, on the other hand, could be wary of it. It remains unclear how facial recognition images will be stored. Could they subsequently be sold to agencies looking to learn more about the spending habits of individuals in order to better target them? We are also unsure how salespeople will act if they receive an alert regarding a potential theft. Will they ask the person to leave and risk falsely accusing a consumer? Vaak has developed highly useful technology that could save retailers billions of dollars. The company must now make sure it protects the security and privacy of consumers, or it could find itself being banned in short order.
By Jean-Michel Vanasse
Collaborator: Jean-Michel Vanasse
Journaliste spécialisé en matière de nouvelles technologies depuis une dizaine d’années, Jean-Michel Vanasse est maintenant animateur et producteur de Planète Techno à ICI Explora et chroniqueur consommation à l’émission Marina diffusée à ICI Radio-Canada.