If AI were a DJ…

If you think cars can fly, you’ll never learn to drive. To use a technology effectively, you must first understand it. That is why the most crucial component in artificial intelligence (AI) literacy is not the practical knowledge of how to use it, but a solid conceptualisation of what it is and what it isn’t. So, how can educators help non-technical students gain this deeper insight?

An old resource can help develop students’ understanding: metaphors. Not just rhetorical tricks, metaphors allow us to make sense of complexity and process the world; in equating one thing with another (“life is a journey”), they highlight certain aspects of reality while obscuring others. Metaphors shape how we perceive and interact with the world. Metaphors matter.

This applies especially to diffuse, complex and evolving concepts like AI. What metaphor best describes this technology? Answering this question isn’t merely a creative exercise; it demands reflection on the essence of AI, our perceptions of it and the directions in which we wish to guide its development.

• Read more about teaching and learning using generative AI
• What does ‘age appropriate’ AI literacy look like in higher education?
• Using AI to foster self-directed learning

A class exercise, used in the Technohumanism: Language, Creativity, and Philosophy of Artificial Intelligence course at IE University, challenges students to come up with more insightful metaphors for AI. It begins with discussion and then students work in teams using big paper sheets and sticky notes to visualise their ideas and offer feedback to other groups.

Here are steps to recreate this experience, along with best practices and lessons learned.

1. Introduce the power of metaphors clearly and engagingly

Begin with a brief introduction to metaphors – what they are, and their importance and purpose. Ask students what they think metaphors achieve. You can reference the classic book Metaphors We Live By to emphasise that metaphors don’t merely reflect reality; they actively shape it.

2. Discuss common AI metaphors critically

Introduce common metaphors for AI and critically discuss what they reveal or distort. Is AI a white humanoid robot, as many images suggest? Or is it rather like Clever Hans, the famous German “mathematical horse”? Is it a “stochastic parrot”, an artificial brain, an “autocomplete on steroids”?

It is worth zooming in on a specific metaphor and its implications in different domains. For example, Melanie Mitchell of the Santa Fe Institute has examined how the idea of large language models as “individual minds” affected the scientific research, use, evaluation and legal regulation of these systems.

3. Address myths and misconceptions about AI

Discuss prevalent myths and misconceptions about AI that could be clarified through better metaphors of this technology. Examples include the mistaken idea that AI has agency and the so-called sociotechnical blindness, the supposed emergent properties of AI, the very term “artificial intelligence”, the anthropomorphic representation of machines, and the reliance on systems being completely accurate and neutral. If this important point has not been covered earlier in the course, good online resources for raising it include the research project AI Myths.

4. Provide a structured outline for metaphor creation

Guide students clearly and practically to design a metaphor for AI. In our case, they completed a form with the following components:

• “AI is…” Clearly state the metaphor.
• “Because…” Explain similarities between AI and the chosen comparison.
• “This helps clarify…” Identify specific myths or misunderstandings the metaphor addresses.
• “For example…” Use the metaphor in a sentence to ensure it fits logically in context.

Together with this outline, give students hints to help them find more interesting and original analogies. For example, recommend that they move beyond comparing AI with a person and explore possibilities such as objects, animals, games or processes.

5. Time to design

Put students into teams of four or five. Each group discusses, designs and writes a metaphor directly on big paper sheets placed around the classroom. This encourages direct discussion without digital distractions. After 25 minutes, they present their metaphor to the class, with the teacher providing feedback. Finally, students walk around the classroom, review the metaphors created, vote for their favourites and add feedback with sticky notes.

6. Wrap-up: main ideas and nuances

Assessing the students’ metaphors reinforces key ideas, with the pedagogical advantage of linking them to visual and therefore persistent images. For instance, one group in our course equated AI with Linguini, the chef from the film Ratatouille who cooks with guidance from the rodent hidden under his hat. This metaphor illustrates that AI is not an alien agent that acts autonomously, even if it seems to, but is designed and guided by humans who should be accountable for the technology’s actions.

Similar metaphors may seem redundant, but they are particularly helpful in identifying nuances. In our experience, different groups compared AI to a mime, a blind sculptor guided by an assistant, a shadow (reformulating Plato’s cave allegory), a cook and a DJ. These may all focus on the same core problem: the relationship between input, output and originality in AI. But what are the subtle differences in each? Which ones best fit the reality of AI?

Designing metaphors for AI can be an engaging activity to promote a creative and critical approach to technology. It encourages students to question dominant views and understand that AI is not a given reality they must accept, but an open future they can shape. Through metaphors, students not only gain a deep analytical insight into AI but also become conscious designers rather than passive consumers of technology.

Pablo Sanguinetti is a professor in the School of Humanities at IE University, Spain.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.