2027 is the year I often think about. AI will be extraordinarily smart by then. Generation alpha will have come up through high school using AI. And I will be 55. I will be done Dear Internet and with academic research. I could move to the teaching track, but I expect traditional teaching will be a mess.

Before then, though, I will continue teaching with two things in mind: process (shorter term) and product (longer term).

First, being able to examine students’ processes will be more important than ever. In this view, multi-staged and scaffolded assignments are the right idea, but the granularity and transparency of a research report or long essay is insufficient. I plan on making use of tools such as Brisk or Grammarly Authorship to be able to follow students’ processes closely and make sure they are adhering to traditional academic integrity norms.

I might even use the Grammarly tool in the fall: it works with Google Docs and Microsoft Word, allows you to play the student’s work, and even assesses how much AI was used. Grammarly is making a useful contribution in this space—Chegg can burn.

Second, in a few years, I don’t think we’ll be able to ask students to deliver products (e.g., research reports or essays) that we traditionally used as proxies/motives for skill development (e.g., research skills and critical thinking). AI can deliver such products in seconds. The products have to be more advanced and the students have to use the AI tools to create them, which itself is a skill that they will need. I don’t know if this will require a change to our notions of academic integrity. And I fear not every student will be able to use AI skillfully to meet the new, higher-level, expectations. I don’t know if traditional education will be able to accommodate this shift.