How artificial intelligence may change our lives

Benedict Cumberbatch plays computer pioneer Alan Turing in the Oscar-nominated film "The Imitation Game." The "game" he speaks of has come to be known as the "Turing Test" for artificial intelligence, or A.I., which has long been a science-fiction staple. Now, it is no longer fiction.

Last summer, for the first time, a computer passed the Turing Test. Scientists are excited, but some people worry about where this all could lead.

Some are already putting these machines to practical use. When professor Manuela Veloso has a guest at her office, she doesn't greet them herself. She sends a robot.

"Hello, I'm here to take Anthony Mason to room 7002," the robot said to CBS News correspondent Anthony Mason. "Please press done when I can go."

More precisely, it's a CoBot, or collaborative robot.

Because the robot is without arms, it needs help pushing the elevator button. But inside the computer science building at Pittsburgh's Carnegie Mellon University, it knows exactly where it's going.

"Hello, I have brought Anthony Mason from room 4405. Please press done when I can leave," the robot said to Veloso.

Veloso and her students first began using the CoBots in 2010. In her building alone, the robot has gone more than 1,000 kilometers.

Four CoBots now roll through the halls. Each navigates on its own computing location and course by using onboard cameras and the detailed maps with which it is programmed.

To send the robot somewhere, you simply hit "schedule task," and the CoBot will ask how it can assist. If you stand in front of it, it will courteously say "please excuse me" until you move.

Unfailingly careful and polite, the CoBot is a determined messenger.

"But as you notice also it does not interact much yet with us," Veloso said. "It's not listening to our conversation, and it could be, but it's not yet. So in some sense, what we need to work on much more, now that it can navigate by itself, is this human-robot interaction."

satmo022215artificialintelligence.jpg
CoBots roaming the halls at Carnegie Mellon
CBS News

Veloso added that they are not far from this goal.

When many of the world's leading artificial intelligence experts met at a conference in Austin, Texas, last month, they wanted to know just how far.

"Turing's test has become an exercise in evasion; it's not really an exercise in artificial intelligence," said Gary Marcus, a professor of psychology at New York University.

"There has been real progress," Marcus said. "But when you think about it the field is 60 years old. And on some of these harder questions, we're still a long way away. Like, having a machine be able to have a real conversation with you. You know, we have Siri. And Siri understands some stuff about movies and restaurants. But it doesn't generally understand what you're talking about. You can't have an open-ended conversation with any of these machines."

Now that scientists can more clearly see the future, they can also more clearly see the risks.

"The biggest danger is when we build programs that we don't really understand that have a lot of control over the world," Marcus said.

British physicist Stephen Hawking sounded that alarm in a recent interview with the BBC.

"Once humans develop artificial intelligence it will take off on its own and redesign itself at an ever-increasing rate," Hawking said.

Tesla founder Elon Musk, at a talk at MIT in October, called it "our greatest existential threat:"

"With artificial intelligence we are summoning the demon," Musk said.

But research is accelerating. Google paid $400 million for the British A.I. company DeepMind last year. And Facebook has set up an artificial intelligence lab, hiring A.I. pioneer Yann Lecun of New York University.

Marcus said that an arms race was on for hiring talent.

"I heard somebody say that whoever dominates A.I. will kind of win the Internet," Marcus said. "So whichever one of these companies really makes that big breakthrough to getting machines that really understand language, for example, is going to have a huge advantage over other companies."

In Pittsburgh, Manuela Veloso sees artificial intelligence as just a giant math problem, one that the CoBots at Carnegie Mellon are beginning to solve.

"My vision is eventually that people will, as soon as they know that this thing exists, they can make use of it," she said.