Is it time to break up with barcodes?

Barcodes have worked well since they were first invented in 1952, but it turns out that they don’t play nice with bots.

The average person can’t read barcodes, but we’ve developed plenty of handy devices that can. And crucially, we excel at getting barcodes in front of those devices. But as robots gradually replace human hands, the time to move on from barcodes may be approaching.

Barcodes are tough for robots because they can be hard to find, especially on irregular objects. If you’ve used a self-checkout lane or worked in retail, you’ve probably struggled to find a barcode at some point. Worse yet, scanning the barcode can become a huge chore if it is crinkled.

Since robots lack sentience, they are all but blind. Cameras attached to robots can collect images, but bots lack the brains to process those visuals. That crinkled label that must be scanned again and again? Good luck getting the bot to recognize the problem, let alone solve it.

Fortunately, humans have big brains, and if the barcode isn’t working, we can manually plug in the code or devise another workaround (paging aisle six!). These days, Amazon aims to teach robots to recognize products visually. If bots and AI can visually identify a product by examining its profile or reading the text on the package, they won’t have to rely on barcodes.

Amazon refers to this system as multi-modal identification, and it’s being slowly developed, rolled out, and tested. “Multi” is currently the keyword, as the system relies on multiple methods to identify products, but developers at Amazon hope that they can make barcodes all but obsolete.