MONTREAL _ The head of Uber’s new self-driving vehicle lab says a viable, on-demand autonomous commercial transportation service remains a long-term goal.
“Having self-driving cars at a smaller scale, on a small set of roads, we are fairly close,” Raquel Urtasun said Tuesday after addressing a Deep Learning Summit in Montreal
“To see at an Uber scale we are far.”
She said much work remains to ensure the technology functions in all possible conditions and locations.
Urtasun declined to predict how far away research being conducted in Toronto will generate the required results.
She said the biggest challenge is the technology itself.
Mapping also remains a very expensive challenge. The cost in the United States alone is estimated at US$2 billion and a cheaper solution is required, she added.
“Nobody has a solution to self-driving cars that is reliable and safe enough to work everywhere,” she said in an interview.
Automotive manufacturers and tech companies are spending considerable money to develop autonomous vehicles.
Yoshua Bengio, an expert in artificial intelligence and head of the Montreal Institute for Learning Algorithms, agrees that it’s going to be many years before vehicles are actually autonomous.
“I think people underestimate how much basic science still needs to be done before these cars or such systems will be able to anticipate the kinds of unusual, dangerous situations that can happen on the road,” he said in an interview.
Urtasun told artificial intelligence colleagues that she chose to work for Uber because she wanted to work in Toronto, not in Silicon Valley, the epicentre of technology in California.
“The Silicon Valley should be in Canada,” she said to loud applause.
“(Also), it is transportation for everybody, not just for the rich. I like that idea.”
Uber has fleets of test cars outfitted with cameras and sensors on the streets of Pittsburgh, Phoenix, San Francisco and Toronto that have travelled more than one million miles.
Urtasun said the goal of her work is to improve transportation safety, increase efficiency, reduce congestion and cut the amount the space used to park vehicles.
“The goal is to get to the transportation of the future.”
Uber Freight is working on developing autonomous vehicles for trucking, which have different requirements than cars used in cities.
Urtasun defended the potential job displacement that would be caused by a commercial driverless Uber fleet, even one that works in concert with a service with drivers.
She noted that disruptions in the past weren’t necessarily bad. She pointed to the impact of ATM machines on tellers and tractors compared to horse-drawn carriages.
“There will be a disruption but hopefully there will also be a lot of other new jobs that will be created as well.”
Bengio was more cautious, noting that the risk of job losses due to artificial intelligence is real, and that politicians should plan accordingly.
“I believe that governments should start thinking right now about how to adapt to this in the next decade, how to change our social safety net to deal with that.”