"Knowledge That Works" is more than a UBalt tagline—it's the driving force behind everything The University of Baltimore stands for and everyone we serve. Through our faculty's insights and expertise, we invite you to explore what that knowledge looks like in action.
In this article, Professor of Accounting Mikhail Pevzner argues for strategic AI governance over what is becoming more like fear-based policy decisions. Learn why he believes directing AI toward societal challenges beats catastrophizing it.
Something strange is happening in the debate over artificial intelligence. Parts of the American political right and left are beginning to sound remarkably similar.
Sen. Josh Hawley has warned about AI replacing American workers and has pushed for
better federal data on AI-related job losses. Sen. Bernie Sanders and Steve Bannon,
who agree on very little, recently appeared at the same Washington event calling for
stronger limits on AI. Meanwhile, the Trump administration has taken a very different
approach, emphasizing rapid AI development and removing regulatory barriers.
The fear about jobs is real. I am very worried about it, especially with three young
children at home. I often joke with a fellow professor that our children might eventually
live on universal basic income. But, as we say in Russian, every joke is only partly
a joke.
My favorite science-fiction series, "The Expanse," presents a bleak version of this
future. On Earth, a large portion of the population lives on basic assistance because
there simply are not enough jobs. People survive, but meaningful work becomes an elusive
privilege. I certainly do not want that kind of society for my children and their
descendants.
Public debate over AI has fixated on two extremes: unregulated acceleration on one
end, and apocalyptic warnings of mass joblessness or human extinction on the other.
But framing the issue this way misses the real policy challenge: how we actually direct
and govern this technology. The latter narrative intensified recently when former
Anthropic and OpenAI researcher Jacob Coxon resigned, warning that the race toward
self-improving AI could prove catastrophic.
However, catastrophizing is not a policy.
Catastrophizing is not a policy.
The more useful question is whether we can influence what AI is used for. If we leave AI investment entirely to private market incentives, companies will understandably pursue the fastest, cheapest returns. Replacing five well-compensated employees with an automated system provides an immediate, measurable shareholder payoff, even while ignoring the broader social cost of worker displacement. Conversely, mapping the human brain, developing better cancer therapies, engineering clean-energy breakthroughs or reducing the cost of space exploration requires longer horizons and less certain payoffs. Yet pursuing these long-term goals yields far greater collective value.
I strongly believe in free markets. They function well within a tolerable degree of
chaos. With a technology as disruptive as AI, however, we must ask whether a significant
portion of that capital should be channeled toward challenges with massive social
returns.
Perhaps we need an organizational model resembling the Manhattan Project for AI. While
the historical analogy is imperfect—we are not fighting a world war, nor do we wish
to—the organizational lesson is sound: The federal government identified an extraordinarily
complex technical challenge and united national labs, universities, scientists and
private industry to solve it.
The Genesis Mission, launched by the federal government, is a promising step in this direction by convening
national laboratories, universities and private companies to apply AI to major scientific
challenges.
Why not establish several large-scale AI missions organized around urgent societal
problems? While the exact list is open for debate, cancer research and neurological
mapping are clear candidates. Clean energy, climate technology and space exploration
belong on the list as well.
The goal is not for government officials to dictate all AI resource allocation; most
development should remain private and decentralized. Rather, the goal is to establish
robust incentives so that top-tier talent and computing infrastructure are also deployed
toward problems whose societal value dwarfs their immediate commercial return.
Higher education has a vital role to play in this effort. Universities house deep,
interdisciplinary expertise across medicine, engineering, computer science, economics,
business and public policy. Academics can help identify high-impact problems, evaluate
competing methodologies and partner with industry and government to implement solutions.
We also need to become far more practical. A colleague in accounting recently pointed
out the limited real-world utility of much academic accounting research. He shared
a simple example: Current AI models still make a surprising number of basic errors
when reading ordinary scanned invoices and recording accounting entries. Why aren't
accounting scholars focusing on solving practical problems like this? Because current
tenure and publication incentives don't reward it. We need to realign those incentives.
Academia has another equally critical duty: helping de-catastrophize the national
conversation.
In cognitive behavioral therapy, catastrophizing occurs when someone takes a legitimate
worry and jumps straight to an apocalyptic worst-case scenario, producing anxiety
rather than action. Therapists work to break that cognitive trap by grounding patients
in reality. Academia should be doing the exact same thing for the public AI debate.
Government, universities and private enterprises have successfully joined forces during
past technological revolutions that carried both immense promise and profound disruption.
We must do so again.
Business schools have a unique perspective to contribute. Every public-private AI
initiative raises core business questions we study every day at the Merrick School of Business: who funds the work, how incentives are structured and how success is measured. Furthermore,
our students will be the managers and accountants deciding whether AI is deployed
merely to cut headcount or to empower workers to be more productive.
In economic terms, the challenge is steering society toward a better equilibrium.
There is simply too much at stake to settle for either chaos or panic.
