​Toyota investing $50M to research cars that can think for themselves

PALO ALTO, Calif. -- At a press conference this morning, automaker Toyota announced that it will be investing $50 million into establishing joint research facilities at Stanford and the Massachusetts Institute of Technology (MIT) to bring artificial intelligence technology to the road and into the home. The five-year collaboration will be coordinated by former Program Director for DARPA's (Defense Research Project Agency) Defense Sciences Office, Professor Gill Pratt.

Semi-autonomous technologies are already gaining prevalence in vehicles today -- from adaptive cruise control that can maintain a safe following distance behind a lead car to lane-departure prevention that can steer a vehicle to stay between road stripes -- but, according to Pratt, getting to this point in advanced driver aid development has been the easy part. The hard part, which Toyota hopes to work out with Stanford and MIT's help, is creating a smart machine that can not only react to stimulus, but can also make complex judgements about its environment and interact naturally and in concert with its human driver.

The automaker's representatives stated during a Q&A following the announcement that it doesn't necessarily equate artificial intelligence with autonomous driving, but that the two technologies are certainly linked. Autonomous cars, for example, will need an AI that can think its way through inaccuracies in map data or interactions with pedestrians. However, a vehicle with a human driver can still be enriched by AI.

Professor Daniela Rus of MIT's Computer Science and Artificial Intelligence Lab (CSAIL) and Professor Fei-Fei Li, director of the Stanford Artificial Intelligence Laboratory (SAIL) spoke in turn of their goals of "human centric" artificial intelligence.

Rus gives the example of a hypothetical vehicle AI that could recognize when the driver is in a foul mood and respond by playing their favorite album or by more closely monitoring that driver for aggressive or distracted behaviors. The future AI could also notice that the driver has forgotten to call their mother, respond with a reminder and even automatically take over with autonomous driving functions for the duration of the call should the driver seem distracted by the discussion. This car would have to be smart enough to know when to automatically intervene, and in some cases, know better than the driver does.

That last bit sounds a bit like 2001: A Space Odyssey's HAL to me, but the automaker assures that, at least for the short term, it's goal of this collaboration isn't to make cars that don't need a driver, but smart machines that can interact more seamlessly and harmoniously with the people inside and outside of the car.

Pratt also alluded that Toyota's upcoming AI and robotics technologies could allow the elderly to maintain their independence, retaining their mobility both inside and outside of the home.

Though cagey about making precise estimates for when we'll see truly autonomous cars on the road, the automaker did state that we should keep our eyes peeled for hits of increased autonomy and artificial intelligence making their way into products within the next few years.