POLITICS: AI and security cameras could stop the shoplifting epidemic
Earlier this year, the New York Police Department shared a grim statistic: Shoplifting in the Big Apple was up 45% from the previous year.
Heyward Donigan, Rite Aid’s CEO, reported that his company took a $5 million revenue hit as a result of shoplifting rising precipitously, making the drugstore giant just one small part of a very big problem: In 2021, for example, shoplifting cost retailers a whopping $100 billion, which is why so many stores in New York City are going belly up rather than dealing with the ongoing theft.
What might we do? The answer may actually be all around us: The CCTV cameras most retailers already have in place – coupled with the newest tool at our disposal, artificial intelligence.
Consider the following: When the NYPD’s Chief of Crime Control Strategies Michael Lipetri briefed the press on 2022’s CompStat numbers, he revealed that 327 repeat offenders accounted for more than 30 percent of the 22,000 shoplifting-related arrests last year.
Now imagine a city-wide AI platform trained specifically to spot these criminals and alert a store whenever they enter the premises.
Or, even more subtly, imagine a system that can scan customers’ hand movements, alerting security whenever someone makes the sort of rapid swipe motion that shoplifters love, but regular shoppers rarely make.
Neither solution is science fiction; both are currently available from a number of promisingstart-ups and are quickly gaining traction in New York.
The former, involving facial recognition software, recently made headlines when Madison Square Garden’s owner, James Dolan, announced that he intended to use the technology to bar people who were suing him from entering the Midtown arena.
But most camera-based AI solutions can be operated with complete anonymization, allowing store owners to gain insight into anything from loss prevention — picking up, for example, on suspicious patterns that might indicate theft — to consumer choice and behavior.
This means bad guys would be identified via their actions, not appearance — to stave off potential biases.
While online retailers have the advantage of collecting reams of data on each person who stumbles into their e-stores, brick-and-mortar retailers have very little way of knowing what their shoppers want. Which, no surprise there, often leads to missteps.
Here’s one glaring example: Often, retailers offer buy-one-and-get-one-free deals to convince consumers to buy more — which works nicely if you’re in the market for a lot of something.
But if you’re simply looking for just one of something, data shows that seeing a buy-one-get-one deal is a likely turn-off.
Why? This is where human psychology comes in: If you didn’t need two tubes of toothpaste, having it offered to you can feel like just too much to deal with.
We want choice, but not too much choice. We want shopping to stay simple.
That’s the sort of insight you can only gain when you observe consumer behavior in real-time, watching the anonymized feed from in-store cameras and seeing shoppers make the choices that they do.
And it’s one that has tremendous value for brick-and-mortar retailers, for whom shelf space is limited and every stocking decision meaningful.
Cameras can play a major role in helping a bruised and battered post-COVID New York City get back on its feet.
They can help us catch the bad guys and also help create good deterrence — as one California chain learned when it unleashed 400-pound robots to patrol its stores, which left shoppers slightly intimidated and on their best behavior.
And, on the positive side, they can help retailers get much, much better about offering shoppers choices that they actually desire.
It’s time, then, that we change the way we think about cameras, which can help us become smarter about everything from security to retail decisions.
True, cameras are blunt tools, but AI is a much finer one.
Together, they can help us reimagine New York as the gorgeous mosaic of safe streets and bustling commerce it was always meant to be.