March 10, 2018

How Amazon Go Revolutionized Grocery Shopping and Why It’s a Threat to Your Privacy

No one likes waiting in line at the cash register in order to pay. For years, supermarkets have been attempting to make your checkout experience faster. They tried to minimize this annoyance by rolling out self-checkout aisles or starting to offer buy online pickup in stores (BOPIS) initiatives.

But what if optimizing the checkout process simply is the wrong approach. Why even have a checkout at all? In January, Amazon Go, a checkout free grocery store opened its doors to the public after more than a year of beta testing the underlying technology. Simply walk inside the store through a subway-like turnstile gate while scanning a code on your phone, pick up the items you want to buy, walk straight out of the store and get charged on your credit card. No human interaction required. This is what Amazon calls its 'Just Walk Out Technology'.

So how does this actually work?

In short, Amazon uses a mixture of Internet of Things (IoT) and artificial intelligence (AI) technology in order to understand which items you are buying. When entering the store through the turnstile gates, which Amazon refers to as the transition area, the customer identifies himself with his smartphone. Inside, there are hundreds of cameras mounted on the ceiling, monitoring every corner from multiple angles.

These cameras make it possible for a computer vision algorithm to follow customers around the store, analyze what items they pick up, put into their bags or return to a shelf. In addition, weight sensors, pressure sensors and load cells built inside every shelf give the system additional feedback. By combining all these different and independent sources of information, the algorithm can tell with extremely high confidence what exactly it is that you are taking.

Due to the concept of analyzing data from several sources, the system seems to be tamper proof. It continues to function even if some sensors or cameras fail, continuing to track you from other angles. It also doesn’t get confused if you place an item on wrong shelf and afterwards pick it up again. In the end, it will still add the correct item to your shopping bill as the system understands what happened. In reality, even with the occasional mistake, Amazon says that the system is still more financially viable thanks to the money saved on cashiers, security guards, and theft.

What are the current challenges and concerns?

One of the main concerns with Amazon’s technology right now is in regard of privacy. Customers are already used to grocery loyalty programs. Grocery stores do not only know your past purchases, they can also infer a lot from your purchasing behavior about you and your family. However, Amazon Go has the potential of bringing this to a new extreme. While the technology does not rely on face tracking, all these cameras may still collect vast amounts of data, such as your movement patterns within the store or how long you look at a particular item before either buying it or returning it to its shelf. Combine this with your purchase history on Amazon’s e-commerce website and it will be possible for Amazon to paint an extremely accurate picture of what you as a person are all about. As of right now, these are all assumptions since Amazon is very quiet about the amount of data they collect, how long they store it and what they are doing with it.

What do you think?

Would you like waiting in line to be a thing of the past and simply be able to grab whatever you need and walk out, thanks to the help of IoT and AI? Tell us in the comments.