Amazon Said That It Needs At Least 10 Years For Robots To Replace Human In Warehouses

Amazon has a massive logistic system, and as technology is simplifying and sparing human of labor-intensive tasks, it cannot turn a blind eye to this trend. The e-commerce giant will surely make use of AI and robotics. However, a specific timeline to which AI can be in charge of everything is still unknown. As Scott Anderson, Amazon’s regional director of operations said, we had to wait at least another decade for that vision to happen. This gives us a clearer idea of what we can expect from automation, even in an environment suitable for robotics appliance like Amazon warehouses.

As we speak, robots are increasing in intelligence and popularity, but they are doing mostly repetitive tasks that developers program them to do. If you want the robots to do something else, there is lots of time and money will need to be poured in the process.  If we want these machines to take care of more of our more sophisticated work, we have to somehow give them the ability to understand the surrounding, which is still in the researching process. Robots nowadays are capable of picking objects up and mimicking some of human’s hand movements, but that requires lots of training with machinery unavailable to be used in business.

It is true that robots can assemble a Tesla car and produce microchips with precision beyond the ability of a person, but in Amazon warehouse, human labor is still widely used as training robots to have some of the workers’ dexterity is too hard of a process to do in a commercial scale.

As deep learning is elevating the way we train AI in the last few decades, robotics can take advantage of this technology too. They are getting the kind of vision and movements closer and closer to a human’s level. As it will be the one who gains a lot from the advancement of robots, Amazon has come up with lots of measures to promote development in this field.

Apart from Amazon, other labs and organizations are advancing fast in robotics. UC Berkeley’s lab is one of the most prominent institutions. Its budget-friendly machines come in the form of two robotic arms with the control of Blue, a central system is highly regarded. Thanks to an advanced vision system powered by AI, it can do sophisticated tasks like folding a towel.

OpenAI is another lab with many breakthroughs. They also make use of AI in a method named reinforcement learning to train robots to do more sophisticated movements, which can be of great help if the machines are used in a warehouse.

Amazon may need ten years to automate its logistic system entirely, but Gap has already had robots in its warehouse. These machines are the creation of Kindred, a startup in San Francisco. These are actually robotic arms partly controlled by human to do picking tasks.

Amazon now has about 125,000 employees working in its 110 warehouses, 50 delivery stations, and 45 sorting centers. Automation is in the process but accounts for just a small portion. As we speak, robots are used limitedly in their warehouses, and the rest still requires the involvement of human. These machines have not yet reached the level of dexterity and precision to cover most of the tasks and training them to do so is too expensive.

The company is using drives, the compact, Roomba-like machines to deliver goods to human employees but they can only follow a preset path.