Colloquium with Dr. James Evans

Bio:
James Evans is the Max Palevsky Professor of Sociology, Director of Knowledge Lab and Founding Faculty Co-Director of Chicago Center Computational Social Science at the University of Chicago, the Santa Fe Institute, and Google. He is also currently Chief Scientist at AI startups Preseen and SAI Labs. Evans' research uses large-scale data, machine learning and generative models to understand how collectives of humans and machines think and what they (can) know. This involves inquiry into the emergence of ideas, shared patterns of reasoning, and processes of attention, communication, agreement, and certainty. Thinking and knowing collectives like science, the Web, and civilization as a whole involve complex networks of diverse human and machine intelligences, collaborating and competing to achieve overlapping aims. Evans' work connects the interaction of these agents with the knowledge they produce and its value for themselves and the system. His work is supported by numerous federal agencies (NSF, NIH, DOD), foundations and philanthropies, has been published in Nature, Science, PNAS, and top social and computer science outlets, and has been covered by global news outlets from the Economist, Atlantic , and New York Times to Le Monde, El Pais, and Die Zeit.

The Singularity Will Not Be Singular: Complexity or Collapse in Collective Intelligence
Thursday, October 29th at 11 am, Varsity Theater WSC, open to faculty and students.

Abstract: Frontier AI does not think alone, and neither do we. This talk argues that intelligence, human or artificial, is collective all the way down, and that the central risk in the coming century is not the rise of a single dominant superintelligence but the collapse of diversity among the many correlated human and machine intelligences we are growing. I show that contemporary reasoning models succeed by generating societies of thought whose competing perspectives underscore their success. Conversations between diverse models are superior at creativity and complex problem solving. But many forces drive contemporary systems toward cognitive and cultural collapse: overlapping language training data and current safety strategies force unintentionally force expressions toward dominant perspectives; long-running model interaction converges on the same ideas; and as human scientists and citizens engage with models performing balance and certainty, they question their cognitive efficacy and forsake their agency. I conclude by exploring architectures and interfaces designed to enhance collective intelligence, human agency and flourishing in the age of smart machines.

Science, Technology, and Careers Advance by Surprise
Second lecture and RSVP lunch, Friday, October 30th at 11 am, Maeser Auditorium
The lecture and RSVP lunch are open to faculty and graduate students in CMS, FHSS, Philosophy and the Honors Program.

Abstract: I discuss the importance of surprise in science, technology, and my own scholarly journey at BYU and beyond. Beginning at BYU, things that mattered most were not things I planned. As an anthropology major, I lived as a tramp in Salt Lake City studying homelessness, founded a hippie performing arts high school where I taught world history in a tree (and lobbied for the experience to count for a chunk of my major), then did a quantitative historical honors thesis on the economy and inequality of United Order communities. I followed my wife to Harvard, wrote business case studies, took a sociology PhD at Stanford, and landed at the University of Chicago, the Santa Fe Institute, and Google studying how science and technologies come to know. Careers, like science and technology, can advance by surprise. In my work building the science of science and innovation, I have shown that the discoveries and inventions that transform fields and industries are the ones that violate expectations most sharply, and that we can measure expectation violation, predict which findings will matter, and design research programs to pursue it. The same logic now applies to the most consequential technology of our time. AI has begun building AI, with agents that propose hypotheses, run experiments, and write the code of their successors, which raises the stakes of designing for surprise rather than consensus. And contemporary GenAI carries a Heisenberg-like uncertainty principle: the more precisely we model a culture or a science to predict it, the more generative agents reshape what they predict, and the less those predictions hold. That coupling sits at the heart of the cascading and existential risks I work on with the Existential Risk Laboratory and work with the World Bank exploring how low- and middle-income countries could use AI to close rather open a growing productivity gap. I close with what all this could mean for students open to surprise and who want to do surprisingly useful work.

Recent, relevant readings: