ITBusiness.ca

Is Canada’s AI potential getting lost in legislative process? Critics weigh in at Data Effect 2023

L to R: Panel Chair Claudia Krywiak, Senator Colin Deacon, Keith Jansa, Mai Mavinkurve during a panel at Data Effect

The Canadian AI and Data Act (AIDA), one of the world’s first pieces of AI legislation, is now in Committee, more than a year after it was tabled, and is mired in classic Canadian legislative brouhaha.

This legislative gap was a key topic during a panel moderated by Claudia Krywiak, president of Ontario Centre of Innovation, at Data Effect: Canada’s AI Future, a half-day event hosted by CityAge in Ottawa this week.

Mai Mavinkurve, managing partner of Prosperity Global Services, explained how artificial intelligence is, in fact, a gift from Canada to the world by the godfathers of AI (Geoff Hinton, Yann LeCun and Yoshua Bengio), but Canada is reaping none of the benefits.

Canada has “tremendous points of excellence”, Senator Colin Deacon added, yet is unable to do a good job fast enough at regulating AI because “legislation gets bogged down too often in partisan debate.”

He stressed, “How do we become agile enough to regulate when Ottawa and other parts of our country regulate in decades, and these technologies are moving in months? I think a lot of that has to evolve.”

The U.S. is moving much faster, Deacon lamented, as he referred to President Biden’s executive order to regulate AI, signed earlier this week.

The U.S., in fact, delivered a very comprehensive AI strategy, Keith Jansa, chief executive of Digital Governance Council added, exploring a number of different tools, like procurement, IP strategy, standards, certification and, simultaneously, involving all stakeholders. Canada, on the other hand, is not bringing the expertise nor leveraging any of these tools, he argued.

“Let’s learn what a Robertson screwdriver does. Let’s learn what a hammer does. Let’s learn what’s actually in that toolbox. Because we need to pull on all those collective levers to cover this responsibly and to effectively have any sort of semblance of competing with Europe and the U.S. and China, for that matter.”

Mavinkurve concurred that, instead of looking at one specific AI application, legislation needs to look at the foundational infrastructure and systems, one of which is data. 

She noted, “Canadian data is something that we have to be very careful about not allowing that to leak out. And it has monetary value. So when we shift our thinking around the fact that data is an asset, we really are going to better manage and govern that data better within our organization.”

Procurement is another important tool that is severely impairing the private sector, Mavinkurve argued.

“We have a huge challenge with purchasing Canadian technology. Why? Why do we have an issue with purchasing Canadian technology? It shouldn’t take that long. We shouldn’t maybe be given preferential treatment, but we need to figure out ways to level that playing field.”

She revealed that when she was trying to raise money for her business, she was told to sell outside of Canada first and then import back to Canada.

“As a country, we’re very good at converting money into research, ” Deacon added. “But we’re not good at converting research into money. And if we don’t start to worry about that, we’re not going to have the money to pay for the research, and our grandkids are not going to be very well off.”

Mai acknowledged that industry/private sector is over indexed on research and under indexed on commercialization, when it should be present every step of the way in legislation.

“We are the ones that are tasked with, ‘OK well, how do we make this practical?’. I’m not going to make money off of a document that says, ‘these are my principles’. I’m going to make money off of software and infrastructure technology that’s going to add value to the economy. So industry absolutely needs to be at the table every step of the way.”

Jansa further argued that any legislation will be outdated by the time it gets enforced, because technology moves too fast. The key, he noted, is to codify certain specific values, like safe AI and then continually evolve the rules and mechanics like standards and certifications.

Canada’s approach to regulating AI, Deacon concluded, should involve the entire government, as well as Canadian innovators, and not just Innovation, Science and Economic Development (ISED).

Exit mobile version