Will robots take over the world?

AI was going to be very influential in our lives in the very near future.

Microsoft's Dave Heiner does not believe that robots will take over Terminator style.

Well, not yet. You can never say never, he admits.

Heiner is strategic policy advisor at Microsoft and on a new committee the computing giant has set up to address ethical issues around the rise of artificial intelligence.

He visited New Zealand and Australia last week.

Some very thoughtful people were concerned about the risks of artificial intelligence (AI), Heiner noted.

Both Tesla founder Elon Musk and astrophysicist Stephen Hawking have called for tighter regulation around AI based on fears that it could evolve too quickly for humans to keep control.

But Heiner believed there were considerably less apocalyptic concerns about AI that were more pressing in the near term.

His work is more focused on how smart computing could change things like the job market, privacy, advertising and health sector.

"It's about looking at the security of these systems - because we know there are bad actors out there, the reliability of these systems - because as smart as these computers are they lack common sense and can make mistakes - and the fairness."

There was a risk that prejudice and discrimination could be built into AI systems unless social factors are accounted for, Heiner said.

Before people got carried away with regulating AI development itself it was important to remember that existing regulations around all those issues were already in place, Heiner said.

So whether it is a discrimination issue or consumer rights AI is going to have to obey the human laws.

"I've never seen [a law] that says it's illegal to do something, unless its you use AI," he said. "You can't say my AI system just selected white men, so I just hired white men. That's not going to be a defence."

It was likely over time that the technology will throw up situations which would require specific new regulation, he said.

"But it is still very early days. We're still trying to figure out: how does this stuff work. How can we explain the benefits to policy makers and make some guidelines."

It was unfortunate that in the mainstream media there was a lot of talk about Terminator type scenarios because the current situation was so far off from anything like that, he said.

"The work that is being done today is something known as narrow AI," he said. "So there is a lot a focus on getting a computer to do a single thing."

Google attracted a lot of attention when its AI programme managed to beat a human player at the board game Go.

"But all that thing does is play Go. It's not as smart as a two-year-old at anything else."

AI was going to be very influential in our lives in the very near future, Heiner said.

"These AI techniques are a way of extending human intelligence," he said. "They are a way of harnessing the power of computers to help us make better decisions."

One very new example was a Microsoft phone app called Seeing AI, Heiner said.

It allowed blind people to point a phone at an object and have it accurately described by the phone's voice.