top of page
  • Writer's pictureChesapeake Group

The AI-based cameras that can spot shoplifters even before they steal

Vaak, a Japanese company, has developed a software that can detect customer’s intention of shoplifting. The software uses an artificial intelligence-based algorithm, which analyses body language and flags customers that seem to show signs consistent with shoplifting behaviour. It can send out an alert to the staff so they can intervene before anything happens.

How does it work? Vaak’s technology uses the live security feeds from the cameras installed in retail stores. It then monitors customers behaviour, looking specifically at their body language. If there are signs that someone is overly nervous, looking around a lot, fidgeting, or generally restless, the system contacts staff via a smartphone app. The staff can then approach the potential shoplifter, engage in a conversation, and offer help just like with any other customer. This intervention will likely reduce the danger of shoplifting as the customer will now feel identified and maybe under watch.

Opportunities for Retail giants to prevent Shoplifting: Amazon’s new cashier-less grocery concept stores, branded Amazon GO, in downtown Seattle have generated a lot of curiosity lately. These stores have cameras and sensors installed which sense when a customer has walked in and when the items are removed from shelves. Also, there are check-in kiosks near the entrance where customers can scan their phone and register their presence via Amazon Prime. Alibaba is also investing in the cashier-less stores model – it is growing its chain of Hema Groceries stores, where customers can go in, scan the products that they want, and get the details regarding that product. However, there still exists a considerable doubt in distinguishing between customers who are willingly picking up items without paying and those who do it accidently. The challenge of shoplifting is quite prominent to everyone in the retail industry.

Recently, NCR, the banking and commerce solutions company, acquired StopLift Checkout Vision Systems, which specialises in developing intelligent computer vision systems that can distinguish between legitimate and fraudulent behaviour at check-out. Vaak’s software takes this space of innovation to next level and is just not restricted to the check-out – it actively monitors customers throughout their presence inside the store. Vaak AI is already being tested across 16 stores in Tokyo. It is expected to globally roll out soon. Interestingly, many retailers do not want their customers to know about the system to maintain the secrecy.

These technologies do face a lot of challenges as well. With new GDPR regulations coming up in Europe and similar compliances likely to come up in other markets, there is a huge task of compliance, which may be tough since these regulations give foremost importance to customer privacy and data protection. There are also studies suggesting that these AI platforms exhibit racist and sexist biases that are inherent in the data they use.

Still in early stages, these technologies do need a lot of improvement. Seeing the interest that they are generating, we are sure to see a lot of innovation happening in this space. For retail industry, which loses around $50bn annually to theft, these technologies can be of great help to shrink losses.


Commenting has been turned off.
bottom of page