Big grocery stores and supermarkets have identified probing problems linked to losing revenues: the spread of shoplifters.
Being cheated into is inevitable. Security guards are protective enough but don’t have the power to identify most of the shoplifters who enter the store. Staff members might miss the actual act since they have to assist customers’ concerns too. CCTV and security cameras prove to be efficient, but at the time when owners finally able to view the scene, shoplifters have already escaped and happily roamed around downtown.
The United States, a first world country with remarkably advance in terms of technology and lifestyle, experiences the same disturbing case. The land of the free revealed that they too are bombarded with the same dilemma, costing its retailers to lose billions of dollars a year. It’s a hard pill to swallow. But, there are too many shoplifting incidents happening per year which inevitably causes companies a significant sum.
However, Japan may help put these shoplifters behind bars, as it announces its newest technological invention.
A Japanese startup has developed an artificial intelligence software that can quickly detect and distinguish shoplifters in the act. The new invention will be a massive advantage to staff members and retail owners, and will exponentially improve the ways to handle shoplifting. In a sense, the technology helps to alert the staffers once the action is on-going, allowing them to butt in and prevent pilferage.
The system is not yet available in the US, but Vaak, the Tokyo-based company responsible for creating the technology, said that it has to be tested in local convenience stores first before launching in various markets worldwide. In other words, Vaak needs to implement the technology on local US stores for it to be of great use to retailers.
Accordingly, Vaak previous tests done in local stores revealed exceptional results, plummeting 77 percent of shoplifting cases. The data acquired by Vaak on its tests details that the technology able to lower incidents significantly than most high-end security cameras have shown in the past years.
‘Vaakeye’ goal is to develop a system which works hand in hand with a store’s surveillance cameras to catch thieves that busy staffers might miss. Its keen developers trained the system by exposing it to television footage for more than 100 hours, showing honest shoppers versus shoplifters.
How it works involves complex data. However, for the public to understand the function of the said AI technology, here’s Vaak CEO Ryo Tanaka detailing the backdrop of the invention. And here’s what he said:
The system identifies suspicious and illegal activities of customers based on the installed hundreds of shoppers’ behavior such as gait, the manner of walking, hand and body movements, facial expressions and even the way they dress.
The AI technology for shoplifters is entirely different from other shoplifting detection technologies. For example, if the technology spots behavior that may be suspicious enough based on the data stored on its system, it will then alert the personnel or the owner via an app. The staff members will have to decide what actions to take to prevent the act and not the machine itself. He or she can either approach the potential shoplifters or ask if that specific people have something they need.
The bottom line is, the system doesn’t label people as shoplifters or thieves because doing so may result in customer grievances and complaints. What the system offers is complete guidance through informing staffers to “please check these individuals, they might steal things” rather than pointing fingers.
Due to the increasing rate of shoplifting in the country, which accounts for billions of losses, Japan’s artificial intelligence has paved the way to combat retail reduction.
Last year, the Japanese communications giant, ‘NTT’ East, made headlines with its invention called AI Guardsman, a camera that uses technology which detects shoplifters through their body languages and gestures. It is the same concept as that of Vaakeye. AI Guardsman’s developers concluded that with its technological cameras, shoplifting was down by almost 40 percent.
However, despite its promising effort to prevent shoplifting, the new AI technology is being criticized for its disadvantages. With the technology installed on most retail stores, regular customers may be afraid of entering, knowing that they are being tracked. And, for some retailers, that would eventually hurt retail sales.
There is also an issue with the systems being biased against shoplifter’s clothing. Unless the learning algorithms tend to be very good at training the system, there will be no problem. However, if the training session teaches that those shoplifters who wear dark-colored jackets might be a potential shoplifter, then the system will create a bias interpretation.
But Jerome Williams, a professor and senior administrator at Rutgers University’s Newark campus, praises Vaakeye’s AI technology. He said that the system’s focusing on body language is a right approach, adding that ‘technology should not racially profile but behaviourally profile people.’
The thing is, Vaakeye pledged that the system had been studied for years and the algorithms used to study the technology are well crafted. The shoplifting AI technology is already out in the world, with the system had been applied in around 50 stores in Tokyo area. And it made a promise to be out soon not only in Japan but in different countries worldwide.