Universities need to step up their AGI game

As Sam Altman and others push toward a future where AI changes everything, universities need to decide if they're going to be leaders or bystanders in helping society navigate advanced AI transitions

Image: Imdjourney

In his reflections for the new year, OpenAI CEO Sam Altman boldly predicts that “in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies.”

It’s tempting to dismiss this as hyperbole. But given recent advances in using AI models to simulate reasoning and exert control over external systems, I suspect there’s a reasonable chance he might be right.

If he is — and if AI continues along the path that Altman and OpenAI are forging toward “Superintelligent tools [that] could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity” — we urgently need new understanding and thinking on how humanity will successfully navigate the coming advanced AI transition.

But where this new understanding and thinking will come from, and how we’ll ensure that it truly benefits society, is far from clear.

We’re already seeing AI companies like OpenAI, Anthropic, and Google, dominate the intellectual space around advanced frontier models and their societally responsible development. But responsible as these companies claim to be (and I think they’re trying hard), they still lack the breadth of vision and understanding that’s necessary to succeed here.

Have you checked out the new Modem Futura podcast yet?

Governments (democratically elected ones at least) have the societal mandate help steer AI toward societally beneficial futures — but in most cases they lack the imagination, vision, or agility, to achieve what’s needed.

Research universities, on the other hand, offer some intriguing possibilities that could — if these institutions step up to the plate — be transformative.

Universities have long-prided themselves as being pivotal in supporting positive societal change. They are — if you buy into the rhetoric — leaders in new thinking; bastions of rigorous and evidence-based scholarship; places where where researchers can transcend intellectual and disciplinary boundaries and push against convention; founts of transformative learning and education; and institutions that are committed to the creation of public value.

But they also tend to be mired in tradition, convention, and self preservation. And as such, it’s fair to ask whether they’re up to the challenge of responding at scale and with agility to the unique challenges and opportunities presented by advanced AI.

I’m not convinced they are. But I’m also struggling to identify other types of institutions and organizations that have the mandate and capability to rise to the challenge. And because of this, I think there’s a unique opportunity for universities (research universities in particular) to up their game and play a leadership role in navigating the coming advanced AI transition.

Of course, there are already a number of respected university-based initiatives that are working on parts of the challenge. Stanford HAI (Human-centered Artificial Intelligence) is one that stands out, as does the Leverhulm Center for the Future of Intelligence at the University of Cambridge, and the Center for Governance of AI at the University of Oxford. But these and other initiatives are barely scratching the surface of what is needed to help successfully navigate advanced AI transitions.

If universities are to be leaders rather than bystanders in ensuring human flourishing in an age of AI, there’s an urgent need for bolder and more creative forward-looking initiatives that support research, teaching, thought leadership, and knowledge mobilization, at the intersection of advanced AI and all aspects of what it means to thrive and grow as a species.

It is, of course, tempting to frame such initiatives along the lines of established disciplines, leading to efforts where “AI” is simply tagged on to exiting endeavors (“AI and biology”, “AI and economics”, “AI and the arts” etc.) But this would be a mistake. As well as constraining thinking, it would inevitably lead to AI being seen as just the latest buzzword needed to secure research funding— with little thought give to real-world impact.

Rather, I would suggest that we need to think in much more integrated and domain-agnostic ways about successfully navigating advanced AI transitions.

One approach is to consider three intersecting foci: How advances in AI could impact where we live (from our homes and communities to the environment and the planet as a whole — space even); How they might transform what we do (from discovering new knowledge and insights, to creating value in all its various and diverse forms); And how they potentially affect our understanding of who we are (from how people behave and function as collectives in society, to the most fundamental aspects of how we define and understand ourselves as individuals).

Three intersecting foci for exploring the understanding and insights necessary to successfully navigate advanced AI transitions

These three foci — and the intersections between them — provide a framework within which radical and audacious efforts could be pursued to better-understand the emerging landscape at the nexus of AI, society and the future; and how we might approach developing the insights and means to successfully navigate it.

To my knowledge, no such efforts currently exist. And apart from a handful of university-based programs that are actively exploring the future of AI and society, most institutions seem to be comfortably stuck in bystander mode as they play with the latest AI tools, or stick to disciplinary norms as they nibble at the edges of AI-based capabilities.

Yet to my mind this framework is one of the more promising ways in which universities could leverage their unique capabilities and positions to be leaders of positive societal change around advanced AI — and one that would compliment the development and application of emerging AI technologies.

Of course, implementing this would require substantial investment in boundary-busting research and scholarship, together with creative ways of ensuring this leads to profound and positive societal impacts at scale — we’re talking billions of dollars here. And it would need funding that does not come with strings attached that stifle creative and audacious ideas — the type of funding that’s more usually associated with philanthropists and foundations rather than government agencies.

It would also depend on a willingness to break from conventional institutional constraints and empower researchers to embrace radical creativity without having to bow to conventional metrics of success

And it would rely on attracting and empowering world-class researchers and scholars who comfortably work across domains and excel at breaking the mold of academic conventions as they set out to change the world.

This all sounds like a tall order for institutions that often take years to make the smallest of decisions. But if humanity is to successfully navigate the coming advanced AI transitions and thrive in an AI-dominated future, it’s time for universities to up their AGI game …

… unless, that is, someone can come up with a better approach that doesn’t rely on academic institutions. In which case, it’s probably time for these hallowed bastions of academia to ask what value they actually bring to the table as advanced AI is poised to transform every aspect of our lives.