Artificial intelligence is a trending topic in both the tech and security industries, but how exactly can it help us in the retail loss prevention industry? Like most emerging technology, it has a lot of potential for growth—and a lot of potential for misconceptions.
Artificial intelligence, commonly referred to as AI, refers to the capability of a machine to imitate intelligent human behavior. In practice, artificial intelligence has a very broad meaning, which allows marketers and salespeople to stretch its definition to suit their needs. For those of us in retail, the first thing that might come to mind when we hear the term AI is its applicability to data analytics. More specifically, we want to know what AI can do for asset protection, from predictive analytics to true prescriptive analytics, as mentioned in an earlier article I wrote in LP Magazine Online in April 2018. In this case, AI promises to take the collected data, analyze it with “machine-learning” algorithms, and help the retailer make the right decisions.
Machine learning is what gives computer systems the ability to progressively improve performance on a specific task, or “learn,” without being explicitly programmed to do so. Like all technology, machine learning has some challenges. One problem in a retail environment is that data is often too vague to translate directly into machine learning. Another problem is that the people who create algorithms often don’t have clean data to work with, which could lead them to create an imperfect or biased algorithm.
But AI has a lot more potential beyond being used to crunch the numbers. Last month, a Japanese startup called Vaak developed an artificial intelligence software they claim can catch shoplifters in the act by alerting staff members, so they can prevent thieves from even leaving the store. CEO Ryo Tanaka said his team used 100,000 hours of surveillance data to train the system to detect suspicious activity using more than 100 behavioral aspects, including how people walk, hand movements, facial expressions, and even clothing choices.
Vaak claims that shoplifting losses dropped by 77 percent during a test period in local convenience stores, demonstrating how this technology could help reduce global retail costs from shoplifting, which hit $34 billion in 2017 according to the Global Shrink Index. Furthermore, implementing AI-based shoplifting detection technology would not lead to a significant increase in costs because security cameras, which comprise most of the required hardware, are usually already in place at retail stores.
Vaak’s technology demonstrates how artificial intelligence can work with facial recognition software, which scans “faceprints,” a code unique to an individual, just like fingerprints. Unlike fingerprints, faceprints can be scanned from a distance, which opens the possibilities of facial recognition’s applications in fields such as security and law enforcement. According to a December 2018 Forbes article, several local public security bureaus in China have started implementing the use of augmented reality glasses, created by the Xloong company, which are able to cross-reference faces against the national database to spot criminals.
This isn’t the first time AI has been used to fight retail shrinkage. Last summer, another Japanese company, the communications giant NTT East, launched AI Guardsman, a camera that uses similar technology to analyze shoppers’ body language for signs of possible theft. AI Guardsman’s developers said the camera cut shoplifting by 40 percent.
Installing artificial intelligence and facial-recognition software does raise some questions about the ethics of the technology, especially when it comes to customer consent. Customers are typically willing to sacrifice some privacy for convenience when they are aware the technology is being used. Most retail stores already post signs about the presence of security cameras, so resolving this concern could be as simple as adding a notice about facial recognition to these signs.
Despite how far science has come, AI does not truly think like a human being just yet. This could lead to a bias in a system’s algorithm. However, just as artificial intelligence can be inadvertently given a bias, it has the potential to be less biased than a human being. This is simply a case of auditing the algorithms to root out any potential bias before training the artificial-intelligence system.
Artificial intelligence in retail isn’t a hypothetical anymore. Today, AI algorithms run inventory management, delivery optimization, and customer-support chatbots on websites, which we are all too familiar with. When paired with facial recognition software, artificial intelligence can even eliminate the need of salespeople, best shown in Amazon’s self-service brick-and-mortar stores that use image and video sensors to shape the customer experience. With artificial intelligence entering the retail loss prevention sphere, we’re going to see great change in how our departments catch shoplifters and combat retail shrinkage.