What is Collective Intelligence?
The question is on my mind because I recently spent a week in Santa Fe, New Mexico with a group of people from around the globe who were curious enough about the topic to leave their homes in search of an answer. Our group was skillfully convened by the Santa Fe Institute faculty and staff who arranged for a three-day inaugural symposium and short course on the topic. If you haven’t interacted with the Santa Fe Institute yet, then follow this link to learn more about their rich educational offerings. Highly recommended!
The collective intelligence symposium and short course program integrated presentations from experts in the disciplines of physics, biology, biological physics, economics, engineering, computational science, and psychology with interdisciplinary panel discussions on the presentation topics. Connecting the speakers’ disciplinary specifics was a transdisciplinary orientation in complexity science, and woven throughout the three days were discussions about artificial intelligence particularly in the form of large language models.
Over three days a set of gifted researchers reviewed what is known about collective activity among fish and fireflies and insects and economies and cities and computer operations. The researchers provided insights into how collectivity arises, mostly in species other than humans. Some ideas about collectivity and collective intelligence are distilled below (with attribution in parentheses).
- Adaptive behavior in groups allows for forecasting, anticipating, responding (Flack)
- Behavior is intelligent if it is adaptive, supports decision-making, and solves problems (Theraulaz)
- The underpinnings of collective intelligence are communication, coordination, and consensus (Krakauer)
Toward the end of the symposium and short course the conveners asked participants to discuss definitions of collective intelligence in small groups and then share their ideas with the larger group. One of my tablemates (more on tablemates below) provided a thoughtful and well defined summary of Informational Free Energy Principle to our small group, which he then shared with the larger group. The idea of free energy was new to me, and so I did some reading to get a deeper connection here and here. I ’m not prepared to summarize the principle, its incorporation of cellular brain function, or its mathematization of entropy based on Bayesian and information theories, but I will offer a lay interpretation of the overarching idea after first reading (feel free to comment and improve my mental model). Humans (and other species) work with probabilistic mental models of their environment to minimize unexpected events and maximize reward. Such probabilistic mental models are most likely built from the interaction of genetically predisposed as well as learned mental operations involving perception, attention, and salience. Intelligence (though not formally defined in the articles linked here) must then relate to the fitness of mental operations for minimizing surprise and maximizing reward.
But what of collective human intelligence?
That’s the great thing about being a human in a human group using forms and artifacts of mental operations to think about and talk about collective intelligence. It’s just so meta.
Let’s avoid the meta for a moment and focus on the material.
You know the situation. Approximately one hundred living, breathing humans sat in a hotel banquet room for eight hours a day for three days. You know the set-up. You’ve seen it before at academic conferences (and weddings, oddly). Twenty or more round 8-top tables dispersed around the banquet room, a dais at the front of the room for speakers (and other honored guests, like a wedding party, oddly), water stations (Santa Fe is a high desert, but these water stations are also present at conferences that take place close to sea level). The round tables require that tablemates congregate on one side of the round or turn away from the table (excluding meals) so they might face forward to view the speakers at the dais and the visualization of what is being spoken projected on a screen next to the dais.
Close your eyes. You can conjure this image easily.
We do this as human academics. We share information in typical conference settings. I’ve done it many times. I’ve planned the same type of event, asking others to sit at round tables (it’s what the hotel offers) even though I know that at least half of the tablemates will turn their chairs away from the table to face the speakers….even though I know that sitting for long periods is an invitation to distraction and disengagement … even though I know that the seating and lighting and pacing of the event are not ideal for information sharing. It’s what we do.
But do we have to continue doing it? What else is possible? What interactive experiences might address and enhance what we know and do about collective intelligence?
I raise these questions here not as a criticism of the inaugural collective intelligence symposium and short course. I think the world of the Santa Fe Institute, the people affiliated with the institute, the quality of their work (including event planning), and the resulting possibilities for learning and adapting in our society. I raise these questions because I believe that the folks at the Santa Fe Institute are exactly the right folks to provide innovative, scientifically grounded, and humanistic responses to the questions.
Looking forward to the possibility of an Annual Symposium & Short Course on Collective Intelligences.

Comments
Post a Comment