His cuddly approach aside, Wolven is not
entirely wrong.
Quote:
Originally Posted by Wolven
Yes, I agree. Which is why I didn't like the vibe of Tony's approach to the AI conversation.
Canada needs to get in front of AI and address it head on and I think there are two key issues that need to be addressed quickly.
1) Nationalize AI. If Canadians are going to adopt AI, we need to ensure that the AI is Canadian. It can still be "Microsoft" or "Google" or whatever but they need to house it on Canadian servers and that the data from the AI is not being given to foreign governments or agencies. (This would help ensure local jobs)
Following that, align the AI to Canadian requirements and values to ensure it is not telling people to kill themselves and offering to write the suicide note. It would also be good to force the AI companies to force more fact based results and discourage misinformation.
Any AI company that refuses to play within the rules of Canada gets blocked from being deployed. This wouldn't really come as a surprise to most SaaS companies as many Canadian companies force their data to be in Canada data centres to protect the data from US Government overreach (they have a law saying they can look at anything they want in a US data centre).
2) Catch the fallout of AI on the workforce. We already have SOX reporting and SOX compliance. The audits can cover things like "how many computers did a company buy and how many are in use". All that would really need to be done is expand the SOX audits to include AI consumption and AI agents. Once you have that data you can ensure the proper taxes are applied based on their use of AI to replace the workforce. |
"Nationalize AI" isn't the right turn of phrase, it's "data sovereignty", and Canada has a serious problem there. We run on foreign cloud infrastructure owned by American companies, and even when the servers sit in Montreal or Toronto, the parent companies fall under US law. The CLOUD Act lets the US compel access to that data. Canada has no sovereign cloud equivalent and no domestic provider big enough to replace the American providers, so we keep getting more dependent on them. Blocking new AI deployments is a fantasy unless we want to kneecap ourselves. We are not the EU, and we do not have that kind of market leverage.
PIPEDA is outdated, we have nothing like GDPR, and the federal government hasn't built a coherent baseline for cross-border data control or AI-processed personal data. That is a national issue for
any government to solve, not a differentiator the NDP can build a revival around. Every party
should fix it, but none will
win votes on it.
The NDP is in a rebuild, they are not forming opposition anytime soon, and they are not going to win on tech policy. If the party wants a reason to exist again, it really should go all-in on labour. That means hitting automation and AI's
impact on workers, not drifting into jurisdiction fights they can't control anyway.