Behind the Chatbot: A Flipped Classroom Exploration of AI Mechanics and Ethics

This flipped classroom activity walks students through the mechanics of generative AI by combining an interactive pre-class AI dialogue with an in-depth workshop on information literacy and critical evaluation.

Part 1: Pre-Class Activity (20–40 minutes)
Students engage directly with an AI chatbot using a structured prompt that guides them through activities, like a role-reversal exercise where they act as the AI to understand next-word prediction, and information. The conversation covers key concepts including training data, probability, vector embeddings ("the geometry of language"), temperature ("the chaos knob"), attention mechanisms ("the highlighter"), and the "Black Box" nature of algorithms. Students are asked to screenshot surprising moments and bring at least one question to class.

Part 2: In-Class Workshop (1+ hour)

The workshop builds on the pre-class activity with five modules:

Process, Authority, and Purpose: Explores the information lifecycle, comparing peer review processes with AI generation, and examines traditional authority versus SEO "Domain Authority."

Lateral Reading & Bias: Teaches lateral reading techniques and identification of bias through loaded language, context, and funding source analysis.

Deep Dive into LLM Mechanics: Reinforces probability vs. truth, demonstrates temperature settings, and discusses RLHF (Reinforcement Learning from Human Feedback).

Ethical Landscape: Facilitates discussion on bias/discrimination, intellectual property, transparency, environmental impact, misinformation/deepfakes, job displacement, privacy, anthropomorphism, and critical thinking habits.

Prompt Engineering: Teaches the four-pillar framework (Role/Context, Task, Constraints, Format) and guides students through collaborative prompt-building for research assignments.

Materials:

Materials include a detailed AI prompt, a sample instruction email to students, a lesson plan, and accompanying presentation slides. The prompt features a playful TA persona and introduces one concept at a time. The activity positions the librarian as the "master mechanic" who helps students verify AI outputs through human expertise and verified databases.

This resource is ideal for one-shot instruction sessions, embedded course partnerships, or standalone AI literacy workshops.

Disclaimer: The included prompt is designed to guide AI chatbot behavior, but AI outputs are probabilistic and cannot be fully controlled. Responses may vary across platforms, model versions, and individual sessions. The author is not responsible for unexpected, inaccurate, or inappropriate content generated when students use the prompt. This variability is, in fact, part of the lesson.