The Risk of AI Overcapacity Is Real

Influential research. Trusted analysis.

In this article first published by Fortune, Susan Ariel Aaronson asks: Could artificial intelligence be the new steel? “In the nineteenth century, government officials came to understand that steel would be essential to both economic growth and national security. Accordingly, several countries devised policies that could not only sustain local production, but also prevent foreign producers from competing in domestic markets.” But, by the 1950s, although these countries were producing too much steel, they kept investing in it, even as demand continued to shrink.

Aaronson notes the huge public investments in AI by governments around the world, investments understandable at the national level but that collectively could lead to overcapacity, a situation with significant risks. “When various governments intervene to create and sustain capacity, as they have done with steel and may now be doing with AI, governments will struggle to address the global spillovers. Policy makers should begin talking about this potential risk.”

In this article originally published in The Globe and Mail, Neil Desai says Canada represents an untapped market for generative AI “hyperscalers,” given the country’s productivity woes and low AI adoption levels. Companies such as OpenAI, Microsoft, Amazon Web Services and Google are “creating comfortability, and perhaps even dependency…incentivizing adoption at scale through discounted initial pricing as well as technical and commercial support in exchange for multi-year commitments.”

But, he warns, these companies’ current business model is “inherently problematic. It could enable short-term productivity gains for organizations but at a potentially great long-term cost that is more than what Canadian executives, board members and cabinet ministers are accounting for.”
Recommended

CBC recently contacted Daniel Araya and Branka Marijan for their thoughts on a newly prepared and long-awaited federal strategy from Canada’s Department of National Defence (DND), which warns that the Canadian military’s approach to AI is fragmented.

This strategy is a start, Marijan told the CBC, but “we really need to know how we’re using AI to aid decision-making, especially in defence applications. This is a context where we need a lot more transparency and oversight, which we don’t usually get.”

Araya commented that “it’s not that we don’t need an infantry, or we don’t need people that can fight with weapons. But the fact is, we’re shifting toward automation, augmentation. We need a different calibre of soliders.…there’s probably room for a complementary system between the fighter on the ground and the augmented software developer that provides a platform that makes the whole thing work.” And the DND will have to careful and deliberate about which private-sector companies to work with, Araya says: “This is about ethics of force, ethics of the application of AI and machine learning.”

Read their full remarks here.

The Digital Policy Hub at CIGI is a collaborative space for emerging scholars and innovative thinkers from the social, natural and applied sciences. Here are the latest working papers from the winter 2024 cohort of Hub fellows:

Andrew Heffernan: “Countering Climate Disinformation in Africa”

Tyler Stevenson: “Potential of Decentralized Autonomous Organizations in Developing Countries”

Maral Niazi: “Universal Convention on Artificial Intelligence for Humanity”

Follow the links on the Hub webpage to learn more about the Hub scholars and their work!

Sep. 9 – 10:00 a.m. EDT (UTC–04:00): We invite you to join us virtually next month to launch Robert Gorwa’s new book, The Politics of Platform Regulation: How Governments Shape Online Content Moderation, which examines the emerging domestic and international politics of online safety.

Following Gorwa’s presentation on the book and his key findings, CIGI Research Directors Tracey Forrest and S. Yash Kalash will facilitate a Q&A session with attendees. Find out more and register here.
Follow us
                         
© 2025 Centre for International Governance Innovation