model literacy workshop
Practice slowing down a confident answer.
The workshop turns AI literacy into short tabletop drills. Each drill starts with a plausible LLM answer, then asks readers to highlight what is sourced, what is inferred, what has a date problem, and what would need a clearer boundary. The goal is not suspicion for its own sake. It is a calm habit of making confidence visible and editable.
The red pencil pass
Circle every sentence that sounds final. Then rewrite it with the conditions that would make it true: date, source, region, model version, dataset, or audience. Readers learn that hedging is not weakness when it preserves accuracy.
The citation forecast
Classify a claim as clear, windy, foggy, or stormy. The exercise helps groups talk about uncertainty without turning the discussion into an argument about whether AI is good or bad in general.