Anita began by asking Keyvan about how he got interested in the conditions that support scientific creativity.

Keyvan explained – with great humility – that he didn’t know what he would be working on when he entered the PhD program at the University of Toronto.   Once there, however, he was exposed to ideas from core faculty including Ajay Agarwal and Sarah Kaplan that accumulated into the foundation of his current research program.

Anita, who was Keyvan’s PhD supervisor, sees remarkable coherence in Keyvan’s agenda despite Keyvan’s claims that he was wandering.    What’s there?   A focus on how research teams form; on how they define boundary-spanning questions; and how public policy can invite, encourage, and support research in a particular domain.  When these three are all deployed together, then we often get creativity at a very deep level.

Keyvan is also quite interested in how we know what we know:  How we measure what is known.  In his work with Sarah Kaplan, he has studied the language that scientists use over time to express their ideas. 

In the end, Keyvan is interested in balance.   He wants to “clean up the mess” of our focus on very specialized topics and bridge across complex domains to look carefully at how real people make real decisions about which scientific projects to work on.

Anita asked Keyvan to generalize about his findings, which he said that he finds quite difficult.

In the end, what Keyvan is interested in is thinking about creativity from the scientist’s point of view.   Scientists are not pursuing only breakthroughs.   Scientists do what they like to do.  The academic science system is somewhat closed, which each of us rewarding others within the community with recognition, prestige, and internal celebrity.   Motivating scientists to do things outside their areas of interest is very difficult.  Keyvan described his own research and that of other scholars who are sociologists of science.   These studies have shown that scientists find ways to do what they want under pretty much any condition.

Given that systems of science – the structures that surround scientists – are so critical for the outcomes of scientific processes, what does Kevyan anticipate regarding AI?

Keyvan now leads the AI Center at LBS where he and colleagues are examining Predictive as well as Generative AI systems.   There is a lot of hype, fascination and interest in AI, but we don’t know a ton about how it will unfold.   In a paper with Bryan Stroube, Keyvan finds that AI is quite persuasive, mainly because it pulls information off the internet that was originally designed to get us to buy or believe something specific.   AI can convince us that obviously wrong answers are right.  It can manipulate us into questioning our own judgement.  The worry is that it might generate a lot of bad science.   On the positive side, however, it can make connections across fields and broaden conversations.