Amazon used Indians to verify purchases at ‘futuristic’ marketplaces; understand

In 2016, Amazon surprised the world by testing a smart minimarket. Using a technology called Just Walk Outcustomers simply entered the stores, picked up the products and left without having to pay on the spot.

To enter, all you had to do was identify yourself using your smartphone. At the end of purchases, when passing through a sensor at the establishment’s door, the balance was automatically deducted from the registered card.

But the company now announced the end of the Just Walk Out service and its replacement with another technology a little less bold. Furthermore, the announcement was accompanied by a complaint: this intelligent system actually depended on human employees to operate correctly.

Keep an eye on your purchases

According to a report on the website The Information, Amazon had a team of around a thousand employees in India reviewing and approving Just Walk Out purchases. In total, around 70% of sales were checked by humans, who analyzed video images from cameras in stores to confirm whether the sensors were read correctly.

These contract employees were not disclosed as part of the system and were described by the company as “associated data professionals for Machine Learning”working alongside an artificial intelligence system.

Just Walk Out should only continue to operate in some Fresh stores in the United Kingdom and minimarkets in the Amazon Go family in the United States.

The Dash Cart has sensors that can be read immediately and even helps you weigh fresh products. (Image: Disclosure/Amazon)Source: Amazon

In other locations, it will be replaced by a barcode reader built into the cart. This feature is called Dash Cart and quickly records each item purchased, as well as checkout without requiring a cashier.

What Amazon says

In a statement sent to Gizmodo, Amazon confirmed the replacement of technologies with the use of smart carts in most establishments. Although, she disputes that the vast majority of purchases were analyzed by humanswithout detailing what the correct percentage would be.

Also according to the brand, the role of human employees was to “continuously improve” the machine learning model used in the system.