Edward Niedermeyer: Toyota bets $1 billion that Tesla is speeding

Self-driving cars, like the one Tesla is trying to build, may save lives. Yet car maker Toyota cautions that a driver lulled into complacency by the illusion of technologically-guaranteed safety is likely to kill someone

It was just a few weeks ago that Tesla sent the Silicon Valley hype machine into overdrive, releasing an autopilot system that ratcheted up expectations about a new world of autonomous-drive technology. Yet almost as soon as the new software feature was beamed into the computers of Tesla's Model S sedans, YouTube videos of daredevils pushing the system's limits showed that the innovation, marketed as a safety device, had unleashed a lot of danger. This week, Chief Executive Officer (CEO) Elon Musk told analysts on the firm's quarterly earnings call that Tesla would have to put "additional constraints" on the system in order to "minimise the possibility of people doing crazy things with it".

There is a lesson here: For all the power of Silicon Valley's narrative of disruptive innovation, it's becoming clear that its mantra of "moving fast and breaking stuff" has very different implications in the realm of the dangerous hardware known as cars than it does for social media apps. And the legacy car makers are taking it to heart.

Perhaps the greatest contrast to Tesla's headlong rush comes from its erstwhile partner in electric-car development, Toyota. At the same time the Japanese company this week announced a staggering $1 billion investment in artificial intelligence research, it tried to tamp down the expectations building around autonomous drive technology.

After the announcement of the new Toyota Research Institute, I talked with Gill Pratt, the man tapped by Toyota CEO Akio Toyoda to run it. Pratt, a former Massachusetts Institute of Technology professor and founder of the Darpa Robotics Challenge, was unequivocal about the technical and social challenges involved with autopiloting. "I think that liability is probably the number one issue," he told me. "These artificial intelligence techniques are very good, but as I explained to Mr Toyoda, the autonomy will not be perfect. It will not prevent all car accidents."

Pratt's warnings reflected both Tesla's autopilot struggles and the huge damage Toyota suffered with supposed 'sudden unintended acceleration' in several of its US models five years ago. In that scandal, Toyota was fined and sued for a nebulous 'defect' whose root cause was never definitively established. Just imagine the potential liability involved with technologies that take nearly all control away from drivers.

Even though he has $1 billion to hire the best minds in robotics and software for his new research and development (R&D) institute, and a trillion miles per year of potential data from Toyota's global fleet, Pratt sees the fully-autonomous driving technology envisioned by Tesla and Google as being too far off to even estimate a deployment date. In addition to working on what Toyota terms 'serial autonomy', the company is also developing 'parallel autonomy' systems, with the car and driver working together as a team.

Pratt isn't the only one concerned over managing the risks inherent to autonomous drive. Lawrence Burns, a former R&D chief at General Motors, has suggested giving car makers an equivalent to the federal National Vaccine Injury Compensation Program, through which the government could support the broad social benefits of reduced road deaths by insuring automakers against autonomous-drive malfunctions. But for now, beta-testing autonomous driving systems on the public, as Tesla is doing, remains a very risky proposition.

It's tempting to see self-driving cars as a purely technological challenge with many inherent social benefits - the most important of which is saving lives. But the autonomous revolution will require a fundamental reassessment of a broad variety of social values, not the least of which involves accepting the inevitability that, at some point, an unpiloted machine or a driver lulled into complacency by the illusion of technologically-guaranteed safety will kill someone.

If this takes place before we have properly prepared the expectations of the public and the legal system, the backlash could severely set back an opportunity to reduce the approximately 33,000 deaths on American roads each year. In light of this risk, the cautious approach of Toyota and the rest of the legacy auto industry seems less like resistance to imminent disruption and more like a sensible caution born of a deep understanding of the automobile's role in society.

© Bloomberg