The AI-enabled future
- By Matt Leonard
- Oct 26, 2016
When the first iPhone launched nine years ago, the power of networked computers could for the first time be accessed with a pocket-sized device. But the change sparked by mobile devices is just an incremental step toward computer-assisted living, according to panelists at Information Technology Industry Council’s Oct. 25 panel on artificial intelligence.
“We’re moving from mobile first to AI first,” Google Public Policy Manager Sarah Holland said. She quoted from Google’s Founder’s Letter: “Over time, the computer itself -- whatever its form factor -- will be an intelligent assistant helping you through your day.”
Murray Campbell, an IBM computer scientist who worked on the company’s Deep Blue and Watson projects, agreed with this assessment. AI was a hot topic in the 1980s, he said, but the results didn’t meet expectations, so funding dried up. Now the funding is back, and the results are better than they were three decades ago. “This time is different,” he said. “It’s actually having an impact this time.”
Panelists made it clear that businesses developing AI technology have a responsibility to get it right for both the customer and their bottom line. Policymakers should refrain from passing laws that would hinder the potential long-term benefits, they added.
Hilary Cain, Toyota's director of technology and innovation, said the policy questions surrounding AI are hard, and companies have had to wrestle with the ethics internally. Companies are making ethical AI systems, she said, because it would be bad for business to do otherwise. Automobile manufacturers, for example, are considering how AI-enabled vehicles would handle ethical decisions about sacrificing a driver to save a crowd of pedestrians.
When humans have to face these ethical dilemmas and tough choices, Campbell said, they are often forgiven -- it’s an accident, after all. “I’m not saying we should have the same amount of forgiveness for AI, but it’s something to think about,” he said.
AI-based technologies will likely make cars safer in the long run, Cain said. More than 90 percent of accidents are caused by human error, and AI systems will do much better, she said. But “getting the decision right isn’t lost on us.”
Matt Leonard is a former reporter for GCN.