AI cameras now target supermarket shoplifting

A Dutch supermarket is using AI cameras to detect shoplifters, after shoplifting-related losses amounted to 100 million euros in 2023.

In the latest approach to curb shoplifting, supermarkets are turning to the use of advanced cameras that use AI to detect when shoppers hide items in jackets, bags, or pockets.

Jumbo, a leading supermarket chain in the Netherlands, recently announced that it may expand its use of this technology to hundreds of its stores throughout the country.

Jumbo said in January that their shoplifting-related losses had amounted to 100 million euros in 2023, and were increasingly quickly.

The most frequently stolen items were meats, coffee, and cosmetics.

A growth market

The AI technology that Jumbo uses is provided by Veesion, a French company whose system is currently used in more than 25 countries, including 110 supermarkets across the Netherlands.

Its software employs gesture recognition technology that has been trained to identify the movements that are typically associated with shoplifting.

The system sends alerts to the manager’s phone, accompanied by a video snippet of the suspected theft.

According to Veesion, retail shoplifting losses can typically amount to about 3% of a store’s annual sales.

Their AI-driven system, they say, offers a significant improvement over traditional surveillance methods, which rely on the vigilance of security personnel and which typically detect less than 5% of shoplifting incidents.

According to Dennis Oude Alink, a store manager at Jumbo’s Hengelo location, the theft of high-value items is a common occurrence.

He recounted an incident where a customer secretly filled his backpack with roast beef and other premium meats in a less monitored section of the store, only to be detected by the system, which then promptly notified the staff.

The Jumbo store in Hengelo currently has 47 cameras, 15 of which are integrated with Veesion’s shoplifting detection software.

False positives

The software occasionally generates false positives, for example in sections of a store where customers are expected to place items like croissants into bags, leading to unwarranted alerts.

The use of this technology represents a positive step towards mitigating the losses incurred from shoplifting, Veesion says.

By focusing on people’s actions rather than their appearance, their smart cameras aim to provide a fair and effective means of theft prevention.