Toyota buckles down on artificial intelligence for safer driving

Toyota is investing $50 million into artificial intelligence research in partnership with MIT and Stanford University.

Toyota is getting serious about embedding artificial intelligence in its cars to cut down on accidents.

On Friday, the Japanese automaker said it’s partnered with Stanford University and the Massachusetts Institute of Technology to create two joint research centers focused on using technology to make driving safer.

Toyota will fork over $50 million over the next five years that will be divided between the two universities to study the use of artificial intelligence. AI is a hot area of computer science related to teaching computers to learn and make decisions like humans can.

Dr. Gill Pratt, a former program manager for the Defense Advanced Research Projects Agency, or DARPA, and leader of its well-known robotics competitions, is joining Toyota to oversee its AI research. He will be based in Silicon Valley.

“We believe this research will transform the future of mobility, improving safety, reducing traffic congestion, and raising quality of life for everyone,” Kiyotaka Ise, a senior managing officer for Toyota, said Friday.

Although Google has been developing its own self-driving cars and pouring millions into AI research, Ise explained that Google’s car projects seem focused on cutting out the need for human drivers. Toyota, in contrast, wants to pair human drivers with more advanced technology that can adapt to dangerous road conditions like snowstorms or thunderstorms.

The point is to help drivers avoid crashes. The challenge is to make a car system smart enough that it can appropriately decide when to take over from a driver who is at risk of a crash.

“[The car] must ensure that it does no harm, not only some of the time, but almost all of the time,” said Pratt.

Pratt explained that current AI technology is still a long way away from letting self-driving cars roam the streets in varied environments beyond Google’s tests. Toyota is “not saying that full autonomy can never be done,” Pratt explained, there’s just a lot more research ahead before then.

“This is not as simple as talking to your Siri or your phone,” said Fei-Fei Li, the director of Stanford’s Artificial Intelligence Lab.

Pratt also said that he does not believe that government regulation could be a bottleneck for Toyota’s AI research, because his team is not yet concentrating on fully autonomous vehicles. Instead, he compared the developments that researchers will be working on to creating anti-lock braking or similar car safety technology that has passed regulatory muster. It’s just that the new safety features will use advances in artificial intelligence, like computer vision, in which computers can recognize objects and react to them.

Ise believes that Toyota’s research into AI will also have a spill over effect and aid the elderly who can lose the ability to drive safely as they age. In Japan, this is especially a problem because of the country’s increasingly elderly population.

Toyota is not the only carmaker looking at using artificial intelligence technology. Companies like Mercedes and Audiare also exploring AI techniques like deep learning to build smarter cars. However, Pratt insisted that most research into using AI technology in cars has focused on “easy cases” like adaptive cruise control in which a car can suddenly brake if its systems detects that a vehicle in front of it has suddenly stopped.

The “hard cases” involve creating an AI that knows better than the driver during difficult road conditions and says, “I’m going to take over right now,” Pratt said. The race to create the best driving system using AI “has actually just begun.”

http://fortune.com/2015/09/04/toyota-artificial-intelligence-car-research/