Amazon is abandoning AI-powered technology that allowed retail customers to simply grab a product and walk out of the store, while their payment transaction was conducted through automation. This decision comes after reports that the "Just Walk Out" system was reliant on over a thousand remote workers.

Initially introduced in 2018, "Just Walk Out" was part of Amazon's broader strategy to revolutionize retail through artificial intelligence (AI). The automated shopping experience was made available in about half of Amazon Fresh stores, and was touted as relying upon a sophisticated network of cameras and sensors to track customers' purchases. 

Jon Jenkins, vice president of Just Walk Out technology, previously described the retail shopping innovation as a combination of computer vision and machine learning that allows the system to track “who takes what and charge them correctly when they walk out.”

In 2023, Jenkins said, 

Just Walk Out technology detects when a shopper’s hand interacts with a product on the shelf. When that happens, machine learning algorithms make sure the correct item is added to the virtual cart—all without any specific knowledge about the person.

The system proved to be slow and expensive to implement and the e-commerce giant, despite its efforts, never achieved full automation and AI integration. Amazon reportedly filled the gap left by the system's limitations with over one thousand workers in India, relying on them to handle tasks such as processing data and issuing receipts. Often, the cheap, outsourced labor caused delays in processing, reportedly taking hours to send back data so customers could get their receipts.

Amazon spokesperson, Nathan Strauss, disputes the characterization, first reported by The Information, that the tech is reliant on humans, instead pointing to the role of workers in machine learning and validation functions.

In a statement Strauss wrote,

The characterization that Just Walk Out technology relies on human reviewers is inaccurate. The primary role of our Machine Learning data associates is to annotate video images, which is necessary for continuously improving the underlying machine learning model powering Just Walk Out technology. Associates may also validate a small minority of shopping visits where our computer vision technology cannot determine with complete confidence an individual’s purchases.

Additionally, despite the advancements in tech, customer privacy concerns arose over Amazon's data collection practices with the Just Walk Out system. The company's extensive gathering of sensitive customer data, including in-store behavior, raised ethical questions and resulted in a class-action lawsuit filed in 2023.

The Surveillance Technology Oversight Project, a consumer advocacy group, sued Amazon alleging that the company quietly sold customer data to Starbucks for financial gain without adequately informing customers of this practice. The case claims that "at the heart of this system" is the unlawful collection, storage, and dissemination of customers' biometric data with third parties. 

Earlier in 2024, Amazon announced that its Ring doorbell subsidiary would no longer release video footage to law enforcement without a warrant. Speculation suggests this move sprouted from concerns about the potential for additional customer privacy lawsuits. 

Between widespread criticism of the technology as both invasive and harmful to job markets, Amazon is shifting its focus to alternative solutions like scanners and screens embedded in shopping carts called Dash Carts. These shopping carts allow customers to scan items as they shop, eliminating the need for a checkout counter. Dash Carts automatically keep track of items added to the cart and also provide a cashier-free checkout process.

For now, the public acceptance of Dash Carts as a Just Walk Out replacement remains unknown in terms of data privacy and customer satisfaction. 

Related:

Amazon Will No Longer Give Ring Doorbell Footage to Law Enforcement Without a Warrant