AI and Education

I’ve always been interested in education and learning, and I’ve spent a lot of time exploring different educational models, methods, and frameworks. I was a mentor for exchange students coming to my home region or going to different countries for 10 years, worked in afterschool programs, was a peer tutor, completed a teaching internship in graduate school, and have been mentoring designers and design-curious people again for the past few years. Early in my career, I also worked in ed tech and explored MOOCs, flipped classrooms, and tools that support teaching and learning in various ways.

As with many other areas, public discourse suggests that AI is going to upend education. I find it worthwhile to consider a few different forms of education under that premise.

At the most foundational level, we have the model of teaching about, which often centers on the memorization of facts and information about certain topics. It is the basic requirement on which any other level has to build.

At the next level, we have teaching to do, which focuses on applicable capabilities and skills that can be executed in specific contexts. This model was fundamental to the industrial revolution, as it prepared people to perform specific tasks in specific contexts, ideally on demand—think of the school bell training people to adopt a “work shift” mindset.

These are the two levels I personally could see being affected the most by AI. Teaching about is something an AI can likely address, and teaching to do can probably also be covered to a certain extent.

On the flip side, if your own abilities are within the realm of having learned about and having learned to do, they may be more vulnerable to being displaced by AI. Consider work that is highly documentable and standardized—the definition of executing specific tasks in specific contexts.

Now consider another level, which is teaching to be. Beyond learning facts and methodologies to perform tasks, this level also includes theoretical frameworks of operation and relevant connections to other fields.

For designers, an example of this is the First Things First Manifesto, which points out that designers need to not only know how to perform the tasks of their craft, but also be mindful of the context in which they operate and the consequences of their actions.

At this level, we are talking about education as enculturation. It includes values that go beyond the specifics of a performance context. It enables thinking about second-order consequences. Some of it also requires teaching more meta-skills that match today’s requirements, such as critical thinking, flexibility, and the ability to learn new skills. This education must provide a mental framework in which to operate and that serves as a basis to evaluate the unknown.

I believe it is at this level that AI cannot yet replace the teaching or learning component, and that operating at this skill level is not yet replaceable.