Every day, in factories around the world, thousands of people spend hours squinting at tiny circuit boards and other electronic components, looking for imperfections. It’s painstaking work, and Andrew Ng, a leading artificial-intelligence expert who’s already spent years helping tech giants Google and Baidu spread AI across their companies, thinks computers can do it better.
Ng, formerly the head of AI for Chinese Internet company Baidu and the creator of the deep-learning Google Brain project, is the CEO of a new startup called Landing.AI that will help companies figure out ways to incorporate AI. Specifically, Landing.AI, which is based in Palo Alto, California, and has existed for only about four months, is working with manufacturers—including Foxconn, the world’s largest contract manufacturer and maker of Apple’s iPhones—to figure out how AI can help with product yield and quality control.
Ng has a history of being ahead of the curve when it comes to finding ways to infuse companies with artificial intelligence. So his latest attempt to use AI is a strong signal that manufacturing might be ripe for transformation.
Ng says he’s interested in manufacturing in particular because it touches so much of our everyday lives—essentially, he sees it as a way to bring a digital transformation to the physical world.
“There are all these decisions that AI, machine learning, can make in a much more systematic way,” he says.
Ng won’t say exactly how Landing.AI’s technology will be rolled out by Foxconn or other manufacturers on manufacturing lines. But he expects it will include visual inspections, and he says his team has developed a learning algorithm that, after being trained on just a few images, can be used to spot imperfections in small electronic components or camera lenses. And he says Landing.AI believes it understands how to use AI to tune the operations of manufacturing machines such as injection-molding machines.
Willy Shih, a Harvard Business School professor who studies manufacturing and technology, says this kind of AI takeover makes sense for visual inspection, especially as companies like Apple cram more and more electronics into a smaller package. Greater component density leads to yield problems, he points out, as things like tiny solder balls between a chip and a circuit board get even tinier but still must be perfect.
And while there are rising concerns about the potential for AI to automate humans out of jobs, Ng says he hopes to help workers get the skills they need to perform the next wave of manufacturing work. Furthermore, he believes there are plenty of tasks that AI will not be able to replicate, like strategic decisions about where to open up a new manufacturing plant.
A Roomba recorded a woman on the toilet. How did screenshots end up on Facebook?
Robot vacuum companies say your images are safe, but a sprawling global supply chain for data from our devices creates risk.
The viral AI avatar app Lensa undressed me—without my consent
My avatars were cartoonishly pornified, while my male colleagues got to be astronauts, explorers, and inventors.
Roomba testers feel misled after intimate images ended up on Facebook
An MIT Technology Review investigation recently revealed how images of a minor and a tester on the toilet ended up on social media. iRobot said it had consent to collect this kind of data from inside homes—but participants say otherwise.
How to spot AI-generated text
The internet is increasingly awash with text written by AI software. We need new tools to detect it.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.