¾ÅÒ»ÐÇ¿ÕÎÞÏÞ

Is it acceptable for a driverless car to deliberately swerve in a way that saves its own passengers yet kills a pedestrian?

That and other ethical dilemmas related to artificial intelligence will be put under the microscope by Kiwi researchers receiving $400,000 in funding from the charitable Law Foundation trust.

With AI technology set to increasingly transform transport, crime prevention and other areas the Otago University research aims to inform public policy over three years.

"New technologies are rapidly transforming the way we live and work, and (this funding) will help ensure that New Zealand's law and policy keeps up with the pace of change," Law Foundation executive director Lynda Hagen said.

Research project leader Colin Gavaghan said the legal, practical and ethical challenges posed by AI technologies, which learn and adapt for themselves, fascinated him.

His example of the driverless car came from a recent Mercedes-Benz announcement that its cars will be programmed to prioritise passengers over pedestrians when an accident is imminent.

He said crime prediction tools also posed ethical dilemmas because while AI might appear objective, its underlying parameters were set by humans.

"This could result in biases being overlooked or even reinforced," Dr Gavaghan said.

"Also, because those parameters are often kept secret for commercial or other reasons, it can be hard to assess the basis for some AI-based decisions.

"This inscrutability might make it harder to challenge those decisions in the way we might challenge a decision made by a judge or a police officer."