What if there was a system that could decide who to consult for a decision in real time? How would the diversity of the available sources affect the information gathered?
Scott E. Page is a professor of management, social science, and complexity at the University of Michigan. He’s also the author of several books including The Model Thinker: What You Need to Know to Make Data Work for You, Complex Adaptive Systems: An Introduction to Computational Models of Social Life, and The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies.
Greg and Scott discuss the importance of diverse models and perspectives in decision-making. Scott also shares insights on the evolving nature of information access and the role of AI in augmenting diversity in team decision-making processes. The conversation covers themes like cognitive diversity, the role of selection and treatment in maintaining diverse perspectives, and the challenge of fostering a healthy organizational culture where diverse ideas can thrive.
*unSILOed Podcast is produced by University FM.*
Episode Quotes:
How do we design institutions for diversity and better decisions with AI
01:01:46: So how do we design, especially now with AI, institutions, organizations, whether they're for-profits, universities, governments, that creates, you know, better people in a way, right? We're so focused on the allocation or the decision that's being made. At the end of the day, the decisions and the allocations are going to be made by those people. So you're getting the outcome, but you're also getting the people. And how do we kind of—and to your point about the treatment—you also want those people to be diverse, right? And you want to allow them and encourage them to be learning new things. In fact, I think you do not want to solve it because you could not solve it, because it would be like social engineering. But I think you want to have some awareness that particular institutional structures and incentives of structures that you're putting in place are not necessarily creating the world you want—or are creating the world you want.
How AI’s power to curate makes culture more important than ever
28:58: We all know AI can know—these are really nice knowledge maps. But the question is: when you start linking people to the knowledge maps and start saying, ‘We can dynamically bring people into the meeting and get their feedback,’ now you’re suddenly curating. I think AI’s ability to curate, to your point, is amazing. But now, let’s pull the culture thing in. How do I not feel like a gadget? How do I not feel like some sort of widget that the AI is using? I think this is where creating the right organizational team culture is going to be really important.
Invisible forces behind organizational design
01:01:30: What comes for free, whether you like it or not, whenever I design an institutional structure and organizational structure, are the norms, the behaviors, the beliefs, the networks—all that other stuff. The kind of dark matter that really matters for society.
Why simple models fail on complex problems
07:59: If you take something like inequality, it is a complex problem, right? Or the environment. It is a complex problem. Models are simple. So there is no way you can explain something complex with something simple. You are kind of explaining a 16 with a three or something. You just cannot. If the problem is this big and your model is this big, you cannot get it all. But if you have a bunch of models in conversation with one another, then I think you can, potentially, reach a deep understanding. You could predict better, right? I think it is a better way to advance science.
How AI can bridge decision gaps across social inequality
58:12: People who come from families who are well socially connected, who have wealth, who are educated—they get good advice on big decisions that maybe other people do not get. And you can go to the internet to get advice, but you are going to get it pointing in a thousand directions. The question is: will there be ways to have—like, will banks, will others—will they develop AI that they say, “Look, okay, we are going to approve this, but why do you not, you know, use this software and go through this process? It may help you think about particular things.”
Show Links:
Recommended Resources:
Guest Profile:
Guest Work:
unSILOed with Greg LaBlanc 
