January 22nd in Seattle marked the inauguration of Amazon Go: the first Amazon store without cash registers. It uses artificial intelligence to manage purchases and directly charge shoppers’ accounts. Let’s look in more detail how it works and what the risks are.
How does Amazon Go know what you buy?
Amazon Go is Amazon’s first store that uses an accurate system of cameras to track purchases without a checkout lane or physical payments. In short, you grab what you need and go. But how does it work? Let’s drill down into the specifics.
First of all, a code is generated by the Amazon Go app, and customers show it to a digital reader before entering the store. At that point, the system of cameras identifies the user and every product they decide to buy, without needing to see barcodes.
The cameras will send the images they capture in real time to machine learning software. This software can monitor your steps, instantly see the products you pick up from the display and charge your account. The location of every product is precisely tracked electronically, and the shelves are equipped with weight sensors. So if you changed your mind on a drink, you can put it back in its place and it will automatically be removed from your virtual cart.
And of course, Amazon Go won’t have cashiers. The few employees in the store are experts in cooking or technology, and their only job will be to help clients choose products.
The shoplifting experiment: thwarted by Amazon Go
With Amazon’s permission, a journalist from the New York Times tried a shoplifting experiment inside the store, to test if the system could really track which products and how many someone takes.
The journalist tried to wrap a drink in a bag to hide it, but the camera’s weren’t fooled. The recognition system saw through the suspicious behavior and his account was charged for the drink. This is the result of refining the beta version and Just Walk Out technology.
Does Amazon Go put our privacy at risk?
Customers’ concern about facial recognition and identity was fairly clear, but it’s really more of a false problem. As Dilip Kumar, technology vice president has stated that it only recognizes a user’s movements, not their face. This means it’s limited to following their movements from one camera to the next.
So maybe the time has come to change the shopping experience once and for all?
"C’è una sola cosa orribile al mondo, un solo peccato imperdonabile: la noia". Oscar Wilde. Dalla redazione di Dotmug non ci annoiamo di certo. Sempre alla ricerca di notizie, condiviamo, twittiamo, instagrammiamo in costante connessione con il mondo digital.