Episode
3. AI Literacy Meets Learning Literacy: A Conversation with Joshua Thorpe
- Podcast
- EdUp Provost
- Published
- Sep 16, 2026
- Duration seconds
- 2389
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/edup-provost-7039442/episodes/3-ai-literacy-meets-learning-literacy-a-conversation-with-joshua-thorpe/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/edup-provost-7039442/3-ai-literacy-meets-learning-literacy-a-conversation-with-joshua-thorpe.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
In this latest episode of EdUp Provost, host Dr. Gregor Thuswaldner welcomes back Joshua Thorpe, a learning designer at the University of Strathclyde and author of the newly expanded second edition of AI for Students, now published by Bloomsbury. The conversation explores the psychological impacts of "AI Vertigo," the shift toward process-based portfolios, the risks of parasocial relationships with chatbots, and the critical need to intertwine AI literacy with traditional learning literacies. Takeaways AI for Students, 2nd Edition: Now published by Bloomsbury and ~30% longer, the revised edition adds more interactive content and critical reflection spaces. ARF Prompting Framework: Thorpe's Aim, Role, Format (ARF) mnemonic encourages active, interactive engagement with AI in education rather than "vending machine" use. Managing AI Vertigo & Slop: Countering AI-driven overwhelm and misinformation requires balanced perspectives (e.g., Cal Newport's podcast) and integrating AI literacy with broader media/information literacy. Process Over Product: AI's availability creates an opportunity to shift assessment toward documenting drafts, prompts, and problem-solving—though frameworks like the AI Assessment Scale must avoid overloading students with complex rules. Parasocial & Cognitive Risks: AI's anthropomorphic design can foster unhealthy emotional attachment, so students need to understand it as stochastic prediction rather than genuine cognition—while still leveraging AI for productive "desirable difficulty."