The World Has a Plan to Rein in AI—but the US Doesn’t Like It

In December 2018, Canada and France announced plans for a new international body to study and steer the effects of artificial intelligence on the world’s people and economies.

Canadian prime minister Justin Trudeau said the International Panel on Artificial Intelligence would be established by the Group of Seven leading western economies and play a role in “addressing some of the ethical concerns we will face in this area.” It was to be modeled on the UN’s Intergovernmental Panel on Climate Change, which helped establish consensus on the world’s climate crisis and recommends possible responses.

Just over a year later, the IPAI has been renamed the Global Partnership on AI, but it still hasn’t quite gotten off the ground. Six of the G7 are on board—with the United States the lone holdout.

Proponents of the idea say it will help governments get up to speed on AI developments and could build international consensus on limiting certain uses of the technology, such as AI projects designed to control citizens or infringe human rights. The White House says the body is unnecessary bureaucracy that threatens to dampen AI development by being overly cautious.

When the project now known as the Global Partnership was revealed in 2018, Canada held the G7’s rotating presidency. France took over in 2019 and kept pushing the project. It said in May that the G7—excepting the US—as well as the EU, India, and New Zealand were interested and would discuss the design of the new organization. In 2020, the G7 presidency is held by the US.

Cédric O, France’s digital affairs minister, raised the question of the Global Partnership’s future in Washington last month with US chief technology officer Michael Kratsios. In a later interview with WIRED, O said “there is a common consensus but for one country.”

O claims to be sympathetic to US concerns the project might slow America’s tech companies, but he warns that without international coordination, unsavory uses of AI could flourish. He notes how China has used facial recognition and other technologies to strengthen its authoritarian security apparatus—prompting US protests and sanctions on Chinese AI companies. “If you don’t want a Chinese model in western countries, for instance to use AI to control your population, then you need to set up some rules that must be common,” O says.

Because advice or principles espoused by the Global Partnership wouldn’t be legally binding, it’s unclear how much the body really could constrain nations’ AI programs. It would also lack the means to regulate private companies. Lynne Parker, US deputy chief technology officer, says the US still worries the group would be too restrictive.

“Our concerns are that the group could be too heavy-handed,” she says. “We believe it’s unethical to hamper and squash down the development of AI technology to the point where you don’t want to use it.”

Parker also says the Global Partnership looks set to duplicate work already underway at the Organization for Economic Cooperation and Development, a separate group of 36 advanced economies. The OECD has established a network of AI experts to advise members on policy, and it has produced a set of AI principles endorsed by more than 40 countries, including the US, that aren’t too restrictive, she says.

France and Canada say the latest plans for the Global Partnership would make it partially supported by the OECD and designed to complement that organization’s other AI programs. Hans Parmar, a spokesperson for Canada’s Department of Innovation, Science, and Economic Development, said the six members of the G7 besides the US, and several other interested countries, are now holding biweekly meetings and aim to launch the Global Partnership in “early 2020.”