As generative AI becomes ubiquitous in higher education, instruction must move students' AI literacy beyond spotting hallucinations to critical interpretation. Students need to understand that generative AI responses that appear singularly authored and authoritative emerge from a statistical blending of training data, platform policies, algorithmic design, and user prompts. This workshop introduces a critical information literacy framework presenting generative AI as a polyphonic, unreliable narrator. Participants will practice four heuristics (authorship mapping, provenance chasing, interrogating the narrator, and narrator/audience switching) with live tools and adapt them to their own instructional contexts. Please bring laptops and be ready for a fun and educational session. Takeaways will include scenario cards, activity templates, and readings.
Participants will be able to: 1. Map at least three socio-technical forces (like training data, fine-tuning, market goals) that influence what generative AI says, how it says it, and to whom. 2. Apply two of the critical reading heuristics to reveal how generative AI privileges certain knowledge, perspectives, and voices while marginalizing others 3. Design or revise one discipline-specific instructional activity that integrates a critical reading heuristic to support students' agency and critical use of AI
