×

Self-driving car startup Drive.ai names former GM exec to board

SAN FRANCISCO, Aug 30 (Reuters) - Silicon Valley self-driving startup Drive.ai has named to its board a former Wall Street investment banker and General Motors Co executive as investor interest in autonomous vehicles intensifies.

Steve Girsky, a former GM vice chairman and Morgan Stanley alum, was named a Drive.ai director on Tuesday. The appointment comes roughly five months after GM bought Cruise Automation, a one-time rival of Drive.ai, for $1 billion. Just earlier this month, Uber acquired another one-time rival, Otto, for an undisclosed sum estimated at close to $700 million.

Drive.ai is one of a handful of startups building fully autonomous systems for cars, a sector that has drawn more attention since the Cruise and Otto deals.

Drive.ai plans to distinguish itself through the team's expertise in robotics and "deep learning," Carol Reiley, the company's president and co-founder, said in an interview.

Girsky, who left GM's board after seven years in April, was instrumental in reviving the Detroit automaker following its 2009 bankruptcy and brings "deep expertise within corporate management and the automotive industry" to Drive.ai, the company said.

The addition of Girsky is a high-profile get for Drive.ai, which was launched last year with an engineering team from Stanford University's Artificial Intelligence lab and has until now stayed under the radar.

Drive.ai wants to build a hardware and software kit powered by artificial intelligence for carmakers. It is already working with unnamed vehicle manufacturers, but has not disclosed a timeline.

Drive.ai said its first market foray will focus on vehicle fleets on consistent routes, whether for delivering packages, ride-sharing or public/private transit.

Drive.ai believes self-driving is best achieved through deep learning, a complex subset of artificial intelligence in which systems are trained by running massive amounts of data through them until they are able to "think" for themselves.

Reiley, a roboticist, told Reuters that deep learning best resembles how humans think. It is more dynamic than strict machine learning - in which a car drives according to rules programmed through algorithms - and better for interpreting "edge cases," unforeseen events that an algorithm cannot solve.

"I view a self-driving car as a robot," Reiley said. "This will be the first social robot that goes into the world."

She also believes Drive.ai's technology can replicate "emotional intelligence," the part of the brain that deals with social behavior and non-verbal communication - the kind of awareness that drivers rely on at a four-way stop.

"It needs to be baked in," said Reiley, who said cars also need a hardware revamp, with targeted lights and sounds to better communicate with other cars and pedestrians.

Drive.ai, she said, envisions a new vehicle language, "enabling them to show intent and interact in complex ways with humans inside and outside the car."

That could mean, for example, a roof-mounted exterior communication system that signals to pedestrians "safe to cross."

Drive.ai closed a $12 million funding round earlier this year with strategic and other partners, including Oriza Ventures and China-based Northern Light Venture Capital.

(Reporting by Alexandria Sage; Editing by Leslie Adler)