Reflections on the AEDU workshop

A few months ago, I gave a talk at the ALife in Education (AEDU) workshop, held online, under the slightly-provocative title Artificial Life Is Everywhere — So Why Is It So Hard to Teach? The talk grew out of some thoughts that had been nagging at me for a while, and in particular since we ran the Societal Outreach Initiatives for Artificial Life workshop at ALIFE 2025 at ALIFE 2025. That is, before we can begin to discuss how to teach artificial life, we probably need to admit that we don’t fully agree on what it is, or who it is for.

The recording of my talk is below. My section runs from roughly 20:20 to 42:30.

Rather than recap the whole talk, I want to use this post to develop, and articulate a bit more, the part that most seemed to resonate most with the audience. That is, a way of decomposing (the teaching of) artificial life that I described in the talk as “building the ALifer.”

The problem, in brief

The talk opened with a tension that I think that anyone who has tried to explain artificial life to an outsider will recognise. On the one hand, and to its strength, the field is incredibly rich and diverse. It is a melting pot of ideas, behaving at once as a domain, a methodology, and a way of thinking, functioning as a kind of sandbox where biology, computer science, physics, philosophy, cognitive science social sciences, mathematics, and so many other disciplines meet and recombine.

On the other hand, that same richness makes the field oddly invisible. Ideas that are foundational to artificial life, e.g. emergence, self-organisation, evolutionary computation, cellular automata, agent-based modelling, are routinely taught and used by people who never once encounter the phrase “artificial life,” who file it away as a subfield of AI, a bag of optimisation tricks, or something that, to quote a colleague, “sounds like science fiction.” To be fair, I don’t entirely disagree with that last point…

Prior to this talk, I spent a few weeks asking colleagues and students if they had ever come across “artificial life” as a field or an idea, and collected a small wall of these reactions for the talk. You can see more of these in the slides I share (below), but a subset includes things like, “I think it’s a subfield of AI”; “it is about genetic algorithms”; “algorithms we can use for video games”; or “I can’t see how CS can create life.”

None of these is wrong, exactly. Each person has grabbed something “true” about the field in a sense, but each thought of a different one, and none of them the whole thing.

The deeper issue, of course, is that we don’t really know what this “whole thing” is. Artificial life has never had a settled definition. It has a family of them, from Langton’s canonical framing of “the study of life-as-it-could-be in order to better understand life-as-we-know-it” through to descriptions as a synthetic branch of theoretical biology, an engineering discipline, or an interdisciplinary domain spanning virtual, mechanical, and chemical substrates. That plurality is intellectually healthy, but pedagogically expensive. Perhaps even useless. As I mentioned in the talk, artificial life is most legible from inside its own community, and far less discoverable to those who aren’t already looking for it. It becomes a sort-of, “if you know, you know” type of field. Or worse, to the outsider, maybe it’s just a weird cult…

So the challenge for teaching isn’t only about the content we’re teaching. It’s about entry points i.e. how the field is framed and made findable at all.

In the talk, I framed this around three things that artificial life struggles with simultaneously: internal coherence (do we agree with each other?), shared understanding (can we make ourselves understood to neighbours?), and external visibility (can anyone find us in the first place?). Then, in an attempt to try to build out of all of these problems, I proposed an approach towards “Building the ALifer” that leans into ALife’s identity of being (as I refer to it) a “domain, mindset, methodology” all at once.

When we consider that each of these can encapsulate a certain entry point into our field (e.g. people coming from the life sciences have a good handle of the domain of ALife; people coming from computer science are well trained in a methodology, and complexity scientists understand the mindset), then “building the ALifer” becomes a matter of systematically moving someone across all three and letting them find their home in the middle. The triangle is thus not a taxonomy of topics but a description of a trajectory through them.

Please feel free to check out the talk above (and the rest of the incredible workshop talks!). If any of this resonates with you, and if you want to contribute, disagree productively, or just compare notes on how you introduce artificial life to newcomers, I’d really like to hear from you. I’m happy to present or discuss these ideas at workshops, seminars, and reading groups, and I’m keen to hear from anyone who has tried to map their own teaching onto a framework like this. You can get in touch here.

And if you already teach or work in this space, the two questions I keep coming back to are the ones I’ll leave you with: what do you do that you’d call artificial life — and, looking at the triangle, what’s missing?

The slides for this talk are available here, shared under a CC BY-NC 4.0 licence.