Ecommerce giant Amazon is implementing artificial intelligence in twelve warehouses. The technology will screen items for damage before being shipped. This will reduce the amount of damaged items sent out and speed up picking and packing.

Currently, the online marketplace’s warehouse workers are responsible for checking items for signs of wear and tear during the order picking process. “Checking for damage can be time consuming given that most items are in fine condition”, said Jeremy Wyatt, director of applied science at Amazon Robotics. “That’s cognitively demanding because obviously you are looking for something that is rare and it is not your primary job.”

‘Less than 1 in 1,000 items are damaged’

According to the company, less than one in 1,000 items it handles is damaged. However, the total number is significant, as it handles about 8 billion packages a year. The ecommerce giant has been automating its warehouses lately. As it struggles to find workers, it recently added robots to its warehouses in the United States.

Amazon is rolling out the system in 12 locations in the US and Europe.

So far, Amazon has implemented the new AI in two of its fulfillment centers. It is now rolling out the system at 10 other locations in North America and Europe. According to the company, the AI is three times as effective at identifying damage as a warehouse worker.

Items are checked during the picking process

The AI was trained with photos of undamaged items, compared with damaged items. That way, it was taught the difference so it can flag a product when it does not look perfect. The items are checked during the picking and packing process.

If something is broken, the item is moved to a worker to check it out.

Goods are picked for individual orders and placed into bins. These move through an imaging station, where they are checked to confirm if the right products were selected. Now, that imaging station will also check whether they are damaged. If something is broken, the bin moves to a worker.