It’s been interesting to see what kinds of conceptual metaphors have been used to describe generative AI. Conceptual metaphors are ways to talk about something by relating it to something else. It’s been interesting because the ways we choose to speak about generative AI in education matter, because those choices impact experiences, expectations, and even outcomes. Early pedagogical agent research for example, identified different roles for agents often positioning them as experts, mentors or motivators. Even the term pedagogical agents carries its own connotations around the kinds of work that such a system will engage in (pedagogical), compared to say conversational agents or intelligent agents.

Below is a small sample. What other metaphors have you seen?

Update: See Dave Cormier’s related post around prompt engineering, or what to call the act of talking to algorithms.