By Bryan Demchinsky
Canadians are world leaders in the development of artificial intelligence (AI), one of the most important areas in science research today. Now, this country’s top minds in the field are launching a campaign to help ensure that their work is used for good and not harm.
Five leaders in AI research, led by the University of Ottawa’s Ian Kerr, have sent an open letter to Prime Minister Justin Trudeau urging that Canada join a global initiative to ban weapons that use AI.
“It is not often that captains of industry, scientists and technologists call for prohibitions on innovation of any sort — let alone an outright ban,” says Kerr, who is Canada Research Chair in ethics, law and technology at uOttawa.
It speaks to the urgency of preventing AI-controlled weapons from being built or used. The stakes are simply too high to do otherwise, Kerr says.
“Delegating life-or-death decisions to machines crosses a fundamental moral line, no matter which side builds or uses them,” he says. “This is not only a fundamental issue of human rights. The decision whether to ban or engage autonomous weapons goes to the core of our humanity.”
No human oversight
Essentially, AI enables a device — a computer, vehicle or robot — to function without human oversight or intervention. In terms of weapons, it means people lose much or all control of target and kill decisions.
It’s important to take the sci-fi element out of this picture, Kerr says. “Most AI experts are not concerned that the weapons will rise up against humans. This is not the Terminator. This is not HAL from 2001: A Space Odyssey. It’s more about the unpredictability and danger in the system in terms of the sheer speed and incomprehensibility of the autonomous AI decision process.”
States haven’t quite developed autonomous weapons yet, but recent reports suggest that companies such as Russian arms manufacturer Kalashnikov are getting close. Until recently, AI-controlled weapons were seen to be decades away. That timeline is now thought to be down to years, Kerr says, and hence the need to act now.
The Canadian government, which has given the three AI institutions heavily represented among the letter signatories $125 million for research, wants artificial intelligence to become a driver of the country’s economy. It is unprecedented that the heads of the network that will most benefit from the money are asking that brakes be put on how their learning is used, Kerr says.
The international AI community hopes to see progress on the call for an autonomous weapons ban at a conference in November at the United Nations. Nineteen countries have already promoted such a ban, and Kerr wants Canada to be the 20th.
The Canadian experts are in friendly competition with counterparts in Australia, who are also urging their government to become that 20th country. Toby Walsh, a professor of artificial intelligence at the University of New South Wales, is coordinating the Australian campaign. In an open letter, 121 leading researchers in the field call on Prime Minister Malcolm Turnbull to take a stand against weaponizing AI.
“We do not have long to act,” they warn. “Once this Pandora’s box is opened, it will be hard to close.”
As a pioneer in the bourgeoning field of AI and robotics law and policy, Kerr had a privileged vantage point from which to enlist the co-signers of the Canadian letter, who head institutions doing technical AI research. With more than 200 signatures already collected from the country’s AI community, all Canadians are now invited to sign the letter and have their say.
Kerr is “delighted to see the technical community collaborate with those in the humanities, and determined to see more such collaboration, given AI’s potential for social transformation.
“Nobody in AI is a Luddite who wants to say we can’t have technology in the military,” he adds. “But the point is that we should not be weaponizing AI without meaningful human control.”
Sign the open letter to Prime Minister Trudeau