Toyota Tu'u Mata AI Le Vaaia I North American Factories

Toyota Motor Amerika i Matu (TMNA
MNA
) is partnering with AI le vaaia to use artificial intelligence to help it make more intelligent decisions in its factories regarding quality, safety and productivity.

The automaker and Austin, Texas-based startup said Wednesday Invisible AI’s computer vision platform will be installed in all 14 TMNA manufacturing locations in North America.

The system essentially puts electronic eyes on every corner of the operation with AI technology to analyze any possible issues.

“If you can’t see problems then you can’t solve them,” said Eric Danziger, Invisible AI co-founder and CEO in an interview. “What we’re giving them is tools to be able to see more of what’s going on inside their facility and they can then go in and correct any problems and any bottlenecks, making sure people are being safe.”

Indeed, Toyota plans to use Invisible AI’s system to improve quality, productivity and safety beyond observations by humans or simple security cameras.

“We observe our employees assembling vehicles to identify inefficiencies and bottlenecks in their standardized work,” said Stephen Brennan, Group Vice President, Vehicle Production Engineering & Manufacturing Production Innovation Center in comments emailed to Forbes.com. “Invisible AI systems will help us increase the frequency and accuracy of process reviews as well as reduce the time needed to find inefficiencies across processes, giving us more time to focus on improvement.”

Wednesday’s announcement marks a milestone in a two-year collaboration between Invisible AI and Toyota’s Manufacturing Project Innovation Center.

The Invisible AI system to be installed at TMNA plants uses 500 edge AI devices with a built in NVIDI
VIDI
NVDA
Dia
A Jetson chipset, 1TB of storage and a high-resolution 3D camera to track all floor activity.

“That allows us to do a ton of processing at the camera, on the device in real time, all the time,” explained Danziger. “What we have is an AI computer vision model that’s running and that’s constantly processing all the video that’s coming in. What that means is we can give you real time information, real time insights, we can process that data incredibly efficiently. If you want to see a shift report you can see that almost as soon as shift ends.”

Danziger points out the video remains in-house on each device, never uploaded to the cloud or exported anywhere else for analysis.

Factory workers are used to the presence of security cameras but the Invisible AI system represents much closer scrutiny of their work and performance.

Toyota’s Stephen Brennan says workers had significant input regarding privacy concerns explaining, “Privacy is a top priority. Toyota production employees are active in developing, implementing and utilizing Invisible AI’s technology to anonymize the videos outside of their immediate production level.”

Indeed, anonymity is actually built-in to the technology according to Prateek Sachdeva, Invisible AI co-founder and COO who pointed out, “Our AI system looks at joints on human body, there’s no facial recognition, nothing that attributes anything to any one person specifically.”

While the Toyota factory workforce is non-union Sachdeva said his company does understand that unions representing workers at other companies looking to install Invisible AI systems might need some convincing about the security of their members’ privacy.

“We’re moving aggressively with union based shops in the U.S. and Europe where there are privacy concerns,” said Sachdeva. “We’re not trying to do anything privacy invasive that would make them look bad.”

The first installation at a Toyota plant will be this year at Toyota Indiana with an initial deployment of 500 edge AI devices. But wider use of the Invisible AI system is something the automaker is seriously considering.

“We are studying Invisible AI’s technology for use cases outside of vehicle assembly, such as safety and ergonomics. The flexibility of the technology is a large part of its appeal,” said Brennan.