Last July, I wrote about Canada taking pride in being among the first countries to develop AI regulations with its Artificial Intelligence and Data Act (AIDA). I commented that being first isn’t necessarily the best, especially when Canadians might lose out on access to innovative technologies. If the choice is between getting AI regulation right, or getting regulation right now, was there really a need for AI regulations right now, at the expense of regulating AI right?
The July post included a reference to where I argued against technology specific legislation.
I noticed that last week, Canada signed the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law.
Recently, Lazar Radic of the International Center for Law & Economics had an interesting piece, writing that “DeepSeek Shows Why Regulators May be Getting AI Wrong”.
He argues that AI is evolving faster than regulation. In December, a number of international competition agencies issued a joint statement warning “that firms with existing market power in digital markets could entrench or extend that power in adjacent AI markets or across ecosystems”. The agencies, representing the US, the EU and the UK, said: “Given the speed and dynamism of AI developments, and learning from our experience with digital markets, we are committed to using our available powers to address any such risks before they become entrenched or irreversible harms.”
The regulators are concerned that only the richest, best funded giants could afford to train AI models. However, we are seeing numerous alternatives emerge, including China’s DeepSeek, and Mistral AI from France. DeepSeek’s simple origins shook the global equity markets in late January, but Tim Shufelt of the Globe and Mail saw a bright side. “This [DeepSeek] could prove to be a technological breakthrough that makes AI more accessible, with the benefits spreading to companies beyond the Magnificent Seven group of tech giants.” He believes more affordable AI technology would be beneficial to a wide range of businesses, large and small.
Radic says that DeepSeek shows that AI development isn’t limited to a few deep-pocketed American firms. It “may be possible with far lower levels of investment than previously imagined.” Further, he writes that “the open-source AI movement more generally is thriving, with thousands of developers collaborating on decentralized AI projects that challenge the idea that only a handful of companies can drive innovation.”
His concern is that aggressive regulatory intervention could freeze a market that is still taking shape, serving to discourage new entrants rather than fostering them. Radic argues this risk “is compounded when intervention is pursued via regulation, rather than the more flexible case-by-case approach of traditional competition law.” Radic warns, “Premature interventions risk locking in artificial monopolies instead of preventing them—turning “monopolistic AI” into a self-fulfilling prophecy.”
DeepSeek demonstrates that AI may emerge to be one of the more competitive fields in technology. With the pace of AI development moving so quickly, regulators may want to consider exercising greater humility before trying to craft solutions for a problem that does not yet exist.