What follows are some reflections in light of news about AI regulation. The thesis is that, while modern AI emerged largely from Canadian-funded research, broader economic benefits to working people are difficult to see on the horizon.

Throughout 2023, a common theme in political commentary in Canada was whether the country is ‘broken.’ While the national mood has been better, a sense of defeat is perhaps a feature of the Canadian identity.

Lament and Faustian bargains

It’s been said that Canadians have an inferiority complex, specifically in relation to the US. Even one of the iconic texts in Canadian political history is a nostalgic shedding of tears for a nation helpless to maintain sovereignty in the face of US hegemony.

In “Lament for a Nation,” George Grant chronicles how, in the early 1960s, the Central Canadian establishment conspired to bring down then-Prime Minister Diefenbaker in light of his refusal to allow US nuclear arms on Canadian soil. The “lament” was the Americanization of Canadian culture; the inability to live outside the hegemony of US capitalism and the technology that emanates from that system (i.e., the military-industrial-academic complex).

Critics might label the work as dated, the reactionary musings of a grump. Arguably, however, a central insight—that technology can homogenize, disempower, and uproot nations and cultures—remains highly relevant today. Who benefits and what is gained in terms of economic growth or ‘progress’ warrants careful consideration.

A ‘threat to humanity’ or idiocy…or both?

AI is the current debated technology, and key policy decisions about its regulation are now being made. While one might be hesitant to make a comparison to existentially threatening technologies like nuclear arms, noteworthy voices are saying that we perhaps should. As Stephen Hawking remarked:

“Success in creating AI would be the biggest event in human history. Unfortunately, it might also be the last, unless we learn how to avoid the risks.”

My own (admittedly limited) exposure to the AI community has left me not only feeling uneasy but also with the sense that there is a lack of understanding even among those who are researching and developing the technology. It also seems like much of AI is automating idiocy. I’m not alone with this sentiment. AI researcher Eliezer Yudkowsky claims the greatest danger of AI “is that people conclude too early that they understand it.” And author Tom Chatfield said in the “brave new world of big data, it’s artificial idiocy we should be looking out for.”

Commentators seem to sway between AI being a ‘threat to the existence of humanity’ and offering some neat tricks on our devices. The latest fad of AI-generated Instagram influencers is perhaps an example of both: an existential threat and idiotic trick.

What is missing from the discussion is a policy vision and plan that gets real about how AI is part of an economy and society that does not accelerate more of the same disturbing trends: declining living standards, low productivity, inequality, etc. Rather than vague techno-optimism, these are topics policymakers and thought leaders seem to be lacking credible responses to.

Industry-driven policy agenda

Recent sector narratives show little indication that Canada’s tech and policy community is thinking beyond dated neoliberal assumptions about regulation and economic growth. Consider the complaints from companies concerned that Canada’s voluntary AI code of conduct could stifle innovation. In the words of Shopify CEO Tobi Lütke:

“We don’t need more referees in Canada. We need more builders. Let other countries regulate while we take the more courageous path and say ‘come build here.’”

One problem with this predictable ‘tech entrepreneur good, regulation bad’ argument is that it distracts from serious industrial or economic policy. This mentality has been effective in driving equity bubbles and making a few people like Lütke billionaires. But there is growing evidence that e-commerce, social media, and other consumer-driven tech has done little for the real economy.

There are indications that investments in computing technologies have not increased productivity. While the late 1990s and early 2000s saw increasing returns to investments in ICTs, for about 2 decades now Solow’s paradox has reemerged. This does not imply there is no potential. But something about the investment and policy mix is off. Perhaps distinctions between industrial/agriculture applications and the culture industries (like Instagram, Meta, etc.) could be made in setting policy priorities. Or maybe more consideration for accessible training and workforce participation is warranted.

As for the regulation that is being put in place, the public has every reason to believe that the process is being driven by corporate interests. The 2023 AI Index report highlights how the AI state-of-the-art is the domain of Big Tech companies that decide how to balance risk and opportunity in the field. A recent media story indicated that Canadian consultations on AI regulations are heavily skewed toward businesses and industry groups. Earlier this month, an opinion piece in MIT Technology Review “AI is owned by Big Tech” points out that regulation often entrenches the power of companies with access and political clout.

Similar skepticism is warranted for industry-led initiatives like ‘ethical AI.’ A recent Stanford report found that workers involved in promoting AI ethics complained of lacking institutional support and being siloed off from other teams. This after Microsoft and Amazon already cut ethical AI jobs. Initiatives like the ethics centre at the Montreal AI research hub might be well-intended, but Big Tech is bankrolling the Montreal AI scene. We would be naive to think that much questioning of the status quo will take place there.

Conclusion

The Canadian federal government has put lots of taxpayer resources into AI research. No doubt, it paid off… at least for those connected to pools of capital from US-based tech giants. Whether it’s part of an economic strategy and industrial policy that benefits working people remains to be seen. Based on history and trends, one has good reason to be skeptical.