Designing the AI Organization

When Demis Hassabis pitched DeepMind to a few venture capitalists back in 2010, the business plan was almost comically audacious. “Step one: Solve intelligence. Step two: Use it to solve everything else,” he recalls in a conversation at Stanford Graduate School of Business with Stanford University President Jonathan Levin. “And people were quite confused. But we really meant it.” Sixteen years later, the “broad arcs” of that plan have gone “unbelievably well,” says Hassabis, a chess prodigy turned video game developer turned neuroscientist turned Nobel Prize-winning AI pioneer. Today he’s on a mission to create “the ultimate tool for science,” building on his decision to give away AlphaFold, the groundbreaking AI system that predicts the structures of proteins. The future, Hassabis says, is just around the corner: “Ten years from now, I think we’ll realize that we were standing in the foothills of the singularity now.”

What if purpose is what drives us to act, but meaning comes from how those actions affect other people? This week on If/Then, we’re sharing a conversation between two Stanford GSB colleagues: Think Fast, Talk Smart host Matt Abrahams and organizational behavior professor Brian Lowery. Recorded live at Stanford’s LEAD Me2We event, Matt and Brian explore the difference between purpose and meaning, why meaning is rooted in our relationships with others, and how the way we show up shapes the people around us. They also discuss sincerity, leadership, and how to give feedback that helps people feel seen rather than judged. It’s a conversation about the ripple effects of our interactions — and what they mean for how we work, lead, and connect with others.

How do you engage effectively across deep disagreement without shutting down the conversation? This week on If/Then, we’re sharing an episode from our colleagues at Stanford Legal, the podcast from Stanford Law School that looks at the cases, questions, and conflicts shaping public life. In a world where confidence is rewarded and humility can feel like a liability, Stanford Law professor Robert MacCoun argues for something radical: fewer unwavering opinions, more critical reflection, and a better way to disagree. On Stanford Legal, MacCoun joins co-hosts Pam Karlan and Diego Zambrano for a conversation about how “habits of mind” borrowed from science can help citizens, lawyers, and policymakers think more clearly, listen more carefully, and build better public debate around difficult questions that don’t have easy answers. Trained as a social psychologist, MacCoun's work sits at the intersection of law, science, and public policy, with decades of research on decision-making, bias, and the social dynamics that shape how evidence is interpreted. In the episode, he draws on his most recent book, Third Millennium Thinking: Creating Sense in a World of Nonsense, co-authored with Nobel Prize–winning physicist Saul Perlmutter and philosopher John Campbell, to explain why probabilistic thinking, intellectual humility, and what he calls an “opinion diet” are essential tools for modern civic life.

“Humans manage to do so much with surprisingly little,” says Douglas Guilbeault, an assistant professor of organizational behavior at Stanford Graduate School of Business. “Whereas AI, by comparison, is doing relatively little, but with so much power, so much compute, so many resources, and by comparison, relatively fewer constraints.” On a bonus episode of the If/Then podcast, Guilbeault describes the implications of his recent work. Although he readily acknowledges that AI is “increasingly able to do quite a lot,” Guilbeault and his colleagues believe they have identified a key principle that distinguishes human intelligence from machine intelligence — and one which illuminates the limitations of machine thinking. Although some researchers and AI boosters believe both humans and AI learn via optimization, Guilbeault and his colleagues have shown that another process more accurately captures how people distill the seemingly infinite complexity of the world and act based on limited information. “You encounter a lot of noise, a lot of chaos, a lot of randomness,” Guilbeault says. “We somehow figure out how to make meaning and establish strong understandings from within that.” What limitations have you encountered in your work with AI? Share your story with us at [email protected]. Related Content: Douglas Guilbeault faculty profile Read "A Simple Threshold Captures the Social Learning of Conventions" here

Chad Jones, a professor of economics at Stanford Graduate School of Business, recently published a paper, “AI and Our Economic Future.” Using more than 100 years of economic data, he modelled several potential AI-infused economic futures we may experience. These include the good (abundance, we never work again), the not-so-bad (business more or less as usual), and the ugly (a superintelligence that turns on us, among other catastrophic options). Cheery stuff, Jones acknowledges, but essential to face. “I think the ability for an AI to do everything on a computer that the best software engineer can do, that seems like it’s either here now or will be here within five years easily,” Jones says. “Hacking the electric grid, hacking the financial system, these kinds of scenarios are things that we definitely have to worry about. The good news is, I think if we get through that, the ability of AI to transform the economy for good, it is really there and present. And, that would be a very great and bright future.” Related Content: Chad Jones faculty profile What’s the Price Tag for Preventing an AI Apocalypse? At What Point Do We Decide AI’s Risks Outweigh Its Promise?