Amazon Upgrades AI in 'Just Walk Out' Tech for Seamless Shopping

Amazon is upgrading the AI-powered technology behind its Just Walk Out system that allows shoppers to pick up items and walk out without going through a checkout.

Amazon announced it’s adding a new multimodal foundation model to the underlying AI technologies powering its Just Walk Out checkout system for retailers.

The multimodal model improves the accuracy of Just Walk Out by incorporating additional variables including lighting conditions and shopper behaviors. The new Just Walk Out AI now analyzes data from sensors and cameras simultaneously, instead of sequentially, combining multiple inputs to enable higher levels of accuracy.

The underlying AI of Just Walk Out no longer needs frequent retraining because the new model features continuous self-learning. The system continues to learn even when store fixtures are moved or merchandise in the store's catalog is changed.

“For retailers, the new AI system makes Just Walk Out faster, easier to deploy and more efficient,” Jon Jenkins, AWS’ vice president for the Just Walk Out technology, wrote in a blog post. “For shoppers, this means worry-free shopping at even more third-party checkout-free stores worldwide.”

Amazon’s decision to improve Just Walk Out comes after it dropped the feature in its Amazon Fresh stores earlier this year, replacing it with self-service smart shopping carts. The frictionless checkout system remains in use in Amazon’s Go and Fresh stores in the U.K.

Related:Edge Computing, Deep Learning Power Amazon Go

Beyond Amazon’s own stores, Just Walk Out has been deployed in more than 170 third-party locations including airports, hospitals and sporting venues including Lumen Field, home of the Seattle Seahawks.

Now Amazon is expanding the feature, confirming it plans to launch more Just Walk Out stores in 2024 “than any year prior, more than doubling the number of third-party stores with the technology this year.”

“As we scale, the system will continue to learn from everyday shopping scenarios and raise the bar for accuracy and convenience, delivering the benefits of AI to retailers and customers around the world,” Jenkins wrote.

Just Walk Out’s underlying technology uses a combination of RGB cameras and computer vision to determine what items shoppers have picked up as they make their way through a store, allowing them to simply leave and their Amazon account gets charged for the items they took.

Launched in 2018, the original version of Just Walk Out required continuous manual retraining and occasionally took time to identify purchases, especially when a shopper's journey was obscured by poor lighting.

Related:Amazon Palm-Scanning Tech to Hit 65 Whole Foods Stores

The new multimodal foundation model is designed to improve customer journeys by minimizing receipt delays making it easier for retailers to use the technology.

“For example, a shopper might pick up and put down multiple varieties of yogurt, in different combinations and as they are doing so, another customer might reach for the same item, or the freezer door could fog up, obscuring the cameras’ view,” Jenkins wrote. “In complex situations like these, the new model can quickly and accurately determine the actual items taken by each shopper.”

This story first appeared in IoT World Today's sister publication AI Business.