Play Live Radio
Next Up:
0:00 0:00
Available On Air Stations

Will Your Self-Driving Car Turn on You? and Other Important Questions

Stanford Center for Internet and Society/flickr

Researchers, scientists, professors and engineers from around the U.S. and the world are in Austin for the 29th annual Conference on Artificial Intelligence. They're here to talk about the latest developments in the field of artificial intelligence and how those developments are affecting human lives.

Some of the field’s prominent names are speaking about AI’s potential impact on the human race. Will robots steal everyone’s jobs? Will machines render humans irrelevant? Or will they rise up and dominate the human species with their superior intelligence?

Not exactly, says Berkeley professor Stuart Russell. He says people – the media, in particular – have blown AI doomsday scenarios out of proportion.

“Usually they will bring out pictures of Terminator robots, or they’ll talk about the Matrix, or Transcendence, other sci-fi movies about AI,” Russell says. “I think take that with about three buckets full of salt.”

Russell says one key aspect those in AI are working on is “value alignment.” It’s important to develop artificial intelligence that has the same goals and motivations as human intelligence, he says. 

“How do we make sure the machines want exactly what we want? That is, are they our perfect servants in the way, in PG Wodehouse’s novels, that Jeeves is the perfect servant to Wooster?” he says. “Wooster is not the sharpest knife in the drawer, and Jeeves anticipates his master’s wishes, cleans everything up when his master makes a mess, and smooths things over.”

Driving is one of the things that’s becoming increasingly automated with the development of AI technology, and Russell believes that cars, in the near future, will essentially be horizontal elevators.

“You get in, press a button to say where you want to go, you get out. And you don’t think about that there’s a whole bunch of engineering and algorithms in the background making sure you get to the right floor. And you wouldn’t dream of going back to the days where you had to have a driver drive your elevator, and try to get it to stop at the right place, and pull back a bit to get out of the door,” he says.

Both Russell and Rice University professor Moshe Vardi are speaking at this year’s conference. Vardi is speaking about how the development of robotics will result in the elimination of the need for human labor in certain job sectors. Many jobs that fall into the middle class – positions that don’t require complex physical or cognitive skills – are becoming automated, he says.

“Information technology is pushing on the middle class jobs, so there’s talk that the middle class is declining, while the high end is thriving,” Vardi says. “There is a lot of fearmongering, and a lot of optimism, and somewhere between these two – technooptimism and technopessimism – what’s the right way to walk between these two ends?”

Russell believes it’s possible to align machine values with those of humans, and the alignment would provide some protection from the doomsday scenarios of science fiction. Perhaps then, value-aligned machines wouldn’t want to intentionally “steal” a human’s job, if the machine were aware that the human in question had a family to take care of.

Both Vardi and Russell believe that those questions are pretty far out on the horizon at this point, but they know the conversation needs to be happening now to guide technology down the right path.

The Conference on Artificial Intelligence runs through Friday. Follow the conference on Twitter here.  

Related Content