01 There is no need to reiterate how we are at a turning point of uncertainties, given that we still do not know what AI will do. To accept this is also to accept the more significant implication, that the current generation of students will be its spearheaders. The future of AI will be determined by their uses, expectations, and ambivalences. Action plan:
Survey cohort as needed. Track one AI frustration and one AI workaround. Share findings without judgment.

02 Consequently, we need to treat the word “AI” like the word “sport” for the immense variety it could accommodate. Much of what we mean then depends on the purpose and parameters: the time period we have, the strategy, the capacity and endurance of the participants, and the end game.

Action plan:
Before any AI discussion, state three things: your time horizon, your role (user/critic/designer), and what winning looks like.

03 There is really no such practice consequently as “AI Education”. It is a limited term soon to be quite use-less. What are we trying to figure out? The trained model? Gaps in identification? User interface? Machinic dialogue? Data provenance? Output verification?

Action plan:
Ask which layer you are interrogating. Do a checklist: model, data gaps, interface, dialogue flow, provenance, output. Always name which one you are examining.

04 There is no way around facing how, even in its most mundane iteration, AI is extractive, because it is about what the human feeds into the machine.

Action plan:
Log key submissions (a question, a correction, a piece of your writing). Ask: what did it take?
Change one variable: shorten your prompt, refuse to correct its error, or switch to a human source for that one task. The machine works for you, not the other way around.

05 The extraction is a cognitive demand: we notice its limitations and will be moved to correct and fine tune it. It is an antagonistic bond.

Action plan:
After correcting an AI output, pause for ten seconds. Ask: did I just work for it or did it work for me?
If it worked for you: Keep the correction. Then, in one sentence, prompt the AI to explain what it learned from your fix. Use that explanation to refine your next prompt.
If you worked for it: Delete your correction. Restart.

06 We should therefore move away from generalizations about how AI will replace humans and discuss how we will be made to work more to perfect the machine.

Action plan:
Each week, identify one task where you spent more time debugging AI than doing the original work. Draw a red line there.

07 This will not happen in a vacuum. The machine will be fed through the questions and apprehensions of a rest-less generation more cynical of their chances to flourish in the given economy. The question consequently is not how they use AI but how AI will be the site of future antagonisms, from the font of new ideas to the vent of slop, cynicism and resentment.

Action plan:
Host a 15-minute “grievance check-in” with peers. Name one way AI made you hopeful and one way it made you bitter. Track both.

08 AI use is intimate. The tactility of digital access and the increased use of voice commands prove this.

Action plan:
Notice the tap, the swipe, the voice you wake, because you cannot refuse what you do not feel yourself doing.

09 The vastness and non-rapport in the use of AI means that our understanding of it can only be explored ultimately as narratives.

Action plan:
Tell more honest stories about the relationship.