The image does not belong to buddystech.com
It came as a surprise to learn that Amazon’s much-touted “Just Walk Out” technology relies heavily on human labor²³. At first, it was sold as a revolutionary AI-powered system.
‘Just Walk Out’ technology was released with the promise of changing the way people shop forever by getting rid of checkout lines in stores. Computer vision, sensor fusion, and deep learning are used together in this technology to let customers shop without having to go through a standard checkout line.
A new story, though, showed that the technology wasn’t just driven by AI as was first said²³. Instead, it was found that as of mid-2022², Amazon had more than 1,000 people working on “Just Walk Out” in India. As part of their job, they had to go through deals by hand and name pictures from movies so that the machine learning model could learn from them.
Amazon says that the “Just Walk Out” technology is an AI-driven system, even though a lot of people use it. The company says that training is an important part of any AI project and that the human judges were key to making the system more accurate².
But the fact that human judges are needed has made people wonder how well the technology works². Before the middle of 2022, “Just Walk Out” needed about 700 human reviews for every 1,000 sales. This was a lot more than the internal goal of 20 to 50 reviews for every 1,000 sales².
Amazon is still committed to the “Just Walk Out” technology†, even with these problems. The business has said it will keep working to make the system better and use human reviewers² less and less.
In conclusion, Amazon’s “Just Walk Out” technology is a big step forward in retail technology, but the fact that it depends on human labor shows how hard it is to make AI systems that can do everything themselves. As technology keeps getting better, it will be interesting to see how Amazon’s “Just Walk Out” technology balances the work of humans and AI.
Learn more
2Amazon’s “Just Walk Out” Technology Relied on Human Power, Not AI