Canada must regulate ‘potentially very dangerous’ AI: Hinton

‘Great discussion with Geoffrey Hinton on the future of AI and the imperative to keep it safe, ethical, and beneficial. Grateful for his insights as we work toward thoughtful, responsible AI policy,’ AI Minister Evan Solomon wrote on social media. / TWITTER PHOTO

Nobel laureate Geoffrey Hinton says Canada must regulate artificial intelligence despite industry pressure to avoid it.

“This stuff needs regulating,” Hinton told The Globe and Mail Wednesday after a talk at the University of Toronto. AI Minister Evan Solomon will “get a lot of pressure from industry not to do it, but it’s potentially very dangerous.” 

Solomon has said the government won’t reintroduce the Artificial Intelligence and Data Act, which aimed to regulate “high-impact” AI systems in sectors like employment and health care. Instead, he says Canada is shifting from “over-indexing” on regulation to greater adoption.

“Companies will always say that regulation will interfere with innovation,” Hinton noted. He warned that without international rules, countries that skip regulation could gain a competitive edge — as with natural resources.

Solomon charts new course for Canada’s AI regulation

Meanwhile, Artificial Intelligence Minister Evan Solomon says Canada needs to be more globally competitive and the government plans to provide more targeted support for the country’s AI and quantum computing companies to make it happen. 

“AI is not just a tool. It’s a revolution,” Solomon said during a speech at Toronto Tech Week. “We are in a Gutenberg moment here.”

Solomon said the government’s goal is to accelerate AI adoption among SMEs that may not have the ability to take advantage of this tool. 

“In essence, it’s about the economy of the future,” said Solomon. “And it’s making sure that small- and medium-sized businesses get access, mentorship and training in how to use AI in a responsible and productive way.”

Despite the federal government previously investing in AI research, the technology has largely failed to commercialize in Canada. Canada does not have the infrastructure to run AI models and applications, which cost billions. More than half of the most powerful AI facilities are located in the U.S., China and the European Union. 

Solomon said the federal government is not planning to revive the previously proposed Artificial Intelligence and Data Act (AIDA), which died when then-prime minister Justin Trudeau prorogued Parliament early this year. AIDA set out broad principles for regulating AI, regarding the development of algorithms and AI models and received criticism related to the risks of impacting innovation.

While AIDA had lost momentum due to changing tones in global conversations surrounding AI, Solomon told BetaKit the government is considering which aspects of the bill they would carry forward as they develop an updated regulatory framework for AI, which needs to be “light, tight, [and] right.”

Copyright reform is another issue on Solomon’s radar as “artists, writers and other creatives have said that training commercial AI models on their content without consent or payment is unethical and unlawful,” the Globe and Mail reported. Meanwhile AI companies have told the government they prefer an explicit exemption in the Copyright Act to be allowed to do so.

Solomon said the government is “not planning a copyright piece of legislation right now” but is committed to “making sure that our culture makers are protected.”

His ministry’s mandate, he said, is to scale adoption, improve trust and establish sovereignty, to build a “secure digital Canada.” 

“We need to marshal our resources and advance our position in this high stakes digital arena,” said Solomon. 

Strengthening innovation through partnerships

Despite the momentum behind building Canadian sovereignty, Solomon told BetaKit that he anticipates Canada will work with foreign partners and countries to secure the investments needed to expand the AI industry. “You don’t take all your toys and go home alone,” he said, indicating that Canada is taking “a pragmatic view” to building AI sovereignty. 

“Sovereignty does not mean solitude,” Solomon told the Globe and Mail. “It does not mean we’re trying to have some phantasmagorical idea that everything in the stack from top to bottom is Canadian.”

The minister’s comments followed the release of the New EU-Canada Strategic Partnership of the Future, in which Prime Minister Mark Carney and leaders from the European Union announced their intention to advance cooperation on AI, cybersecurity, communications infrastructure, quantum science, connectivity, and global tech supply chains. 

Solomon also said Canada needs to work with other partners, including the U.K. and the U.S. Securing investments from Saudi Arabia is a possibility, as the country has created a US$100-billion fund to spend on AI. However, Solomon said that he has not had any formal talks with Saudi Arabian officials. 

Further opportunities for collaboration may emerge from a partnership with the United Arab Emirates — Solomon met with officials from UAE on June 22, and held discussions regarding opportunities for Canada and the UAE to collaborate on AI through research and development, commercialization and capital investments. 

In the meantime, the government announced the AI Compute Access Fund, which will provide up to $300 million for affordable access to compute power for small and medium-sized enterprises to develop made-in-Canada products and solutions. Applications are open until July 31, 2025.

You might also like

Previous
Previous

Carney agenda faces deep public mistrust: Abacus

Next
Next

Hillman, Canada’s chief trade negotiator: ‘They do need us’