People are reluctant to switch robots off because they see them as "somewhat alive", research shows.

Although people know robots are machines, Canterbury University human interface technology laboratory director Christoph Bartneck has found that they see them as living beings.

Bartneck has been studying human-robot interaction for a decade.

His team of academics, doctorate and master's students conducted experiments, including one in which a person was asked to "kill" a robot by switching it off.

"If you think something is alive, then you probably feel bad about ending that life," he said. "What we were looking for was hesitation and to what degree people were willing to switch off a robot.

"We found with robots that interacted with more intelligence, people were much more hesitant to switch them off."

People became embarrassed in front of a robot, Bartneck said.

"A proof of humanity is if people get embarrassed in front of them. We are embarrassed towards other humans," he said.

"We wouldn't undress in front of other people. If you are in front of a computer, it doesn't matter because it's a machine.

"If you are embarrassed in front of a robot, it means there's a social understanding that the robot is alive."

It was important to programme the right amount of humanity in a robot, Bartneck said.

The goal was not to make robots as human as possible.

"If they act more human, you raise the expectation beyond what we can achieve with robots.

"When a robot talks to you, you expect them to understand you, but they don't.

"You have to have the correct robot to interact with humans," he said.

"What we see now are a lot of robots these days entering society. You have vacuum robots and lawn-mowing robots.

"There is a lot of funding for schemes for robots to take care of the elderly.

"Robotics was first introduced to reduce manual labour, and that still holds true for social robots."

- © Fairfax NZ News