Jiwon's Encyclopedia

The Future of Science

I think a lot about what will happen to different scientific disciplines in the future. Which current fads will turn out to be a dead end? What buried paper will be rediscovered and hailed as the forgotten foundation of an entirely new field? How will experimental and theoretical methodologies change and develop?

This article is a place for me to toss out totally crazy predictions about the future of science that will most likely be wrong. Who knows? Maybe I'll get something right one of these days.

Machine Behavior

In 2019, Rahwan et. al. published "Machine behaviour" for Nature that has since received over 370 citations at the time of writing according to Google Scholar. The group of social scientists, computer scientists, and biologists argued for a new paradigm of studying artificial agents, one modeled from animal behavior. (The broad field that encompasses ethology, comparative psychology, behavioral ecology and anthropology.)

In the grand scheme of things, we are at the inception of artificial intelligence as a discipline. It has only really gotten traction in the last two decades. As such, much of the individuals who interact with artificial intelligence are the engineers, mostly with a mathematical and computer science background.

If artificial agents proliferate to all corners of life, however, as The Age of Em predicts, the vast majority of interactions will take place between artificial agents and laypeople as well as in between an ecology of machines. When we reach that point, a reductionistic inside-out approach to study and predict the behavior of artificial agents won't suffice. The emergent behaviors that arise from the interaction of ubiquitous machines and humans could only be understood using experimental methods developed by biologists.

I am sold by the thesis of Machine Behaviour. Already we see think tanks such as OpenAI hire more social scientists for AI research. It wouldn't be too long until ethologists and ecologists get on the boat. I predict that in a few decades machine behavior will become a vibrant, emerging field.


©️ CC-BY Jiwon Chang 2021