Skip to main content

Toyota partners with MIT and Stanford on AI research, poaches head of DARPA Robotics Challenge

Toyota partners with MIT and Stanford on AI research, poaches head of DARPA Robotics Challenge

Share this story

Eric Thayer/Getty Images

At a press conference this afternoon, Toyota announced that it is forming partnerships with MIT and Stanford focused on AI and robotics research, opening joint facilities at both universities and pumping $50 million over five years into the program. Perhaps more interestingly, the company has also hired on Dr. Gill Pratt — the man behind DARPA's Robotics Challenge — to run the new partnerships with the title of "Executive Technical Advisor."

The endgame for these new facilities isn't entirely clear, but Toyota mentions "future mobility" and "improving the human condition" — clearly, there's a lot of technology involved in bringing autonomous driving, so it stands to reason that'll be one of the focal points. "The focus of the effort today is more on the autonomy of people, of human beings," Pratt says, "regardless of the limits imposed by age or illness."

Pratt also notes that his goal is to eliminate highway crashes without eliminating the fun of driving at any age, perhaps suggesting that cars of the future could have a variable level of autonomy and driver assistance features based on the driver's ability level or situational awareness — and it also suggests that Toyota is holding fast against the goal held by some companies of making cars entirely self-driving and steering wheel-free. (Toyota has been reticent to label its autonomy R&D "self-driving" for a long time.) He also mentioned that Toyota wants to research tools for moving goods and people not just outdoors, but indoors as well, allowing humans to "age in place" through the help of robotics.

"It is incredibly fun to drive most of the time."

"I want you to think about the hard cases," Pratt added during a Q&A session, saying that simple safety features like forward-looking pre-collision braking have already been added. More research into AI is needed for a completely safe vehicle, he says, including types that "haven't been invented yet."

Kiyotaka Ise, a Toyota senior managing officer in attendance, noted that the research is "human centric," which is "totally different" from what Google is focused on with its self-driving efforts. "This is like a train without the operator," he said through a translator. "I think it's going to take quite a long time" to get that type of technology commercialized, he said. "It is incredibly fun to drive most of the time," Pratt added. "Why is it such a joy? I think it relates to this innate drive that we have to be autonomous ourselves, and I think back to my own children and watching when they first learned to talk. Before they did so, they were stuck, they relied on us to carry them around," he said. "Can we have it both ways? Can we have the joy of mobility? The joy of agency... and also have the car be safer, more efficient, and even more fun to drive?"

"We are not saying that full autonomy can never be done," Pratt added, but it's not the first step. "I don't think there is a person on Earth who has a legitimate prediction" when fully autonomous cars will be commercialized and ready for full use, he said.