Teaching AI Fluency

Four ways to teach the framework

1. Linear. Delegation, then Description, then Discernment, then Diligence, in order. Best for beginners who need structure.
2. Non-linear. Start anywhere and move flexibly, teaching all four Ds as an interconnected network. Suits experienced students who already have AI habits to hang the framework on.
3. Focused. Deep exploration of a single competency at a time. Best when depth matters more than coverage.
4. Two loops. Teach it as nested processes: the strategic Delegation-Diligence loop containing the tactical Description-Discernment loop.

Match the approach to student readiness, available time, and goals. There is no single right sequence, only a right sequence for this room.

Teaching the outer loop: Delegation and Diligence

The shift you are producing in students: from "how do I use AI?" to "how do I make good decisions about AI?" The Delegation-Diligence loop addresses strategic and ethical decision-making, whether, when, and how to use AI, and owning the results, and it runs both ways: delegation choices raise diligence questions, and diligence realizations reshape delegation.

Teach it through scenarios with decision points, not checklists. A scenario forces the judgment call that a checklist hides, and scenarios help students understand how the competencies connect in practice. One more move: present constraints not as limitations but as creative catalysts; the best delegation thinking happens inside real limits.

Teaching the inner loop: Description and Discernment

The shift here: from commands to conversations. Students typically arrive thinking in prompt tricks; the destination is thinking in working relationships. The target concept is the cognitive environment: a shared vocabulary, established interaction patterns, and mechanisms for building on previous exchanges.

Teaching moves that work: assignments spanning multiple interactions over time, so relationships can actually form; sharing your own AI collaboration process, including the messy parts; and having students document how their interactions evolve.

Assessing the 4Ds

The assessment triad, and all three are needed for the complete picture:

1. Outcome-based: what students produce through AI collaboration.
2. Process-based: how students work with AI over time.
3. Reflection-based: metacognitive awareness, how students think about their own thinking.

Two principles hold it together: observable actions and concrete artifacts are more reliable than assumptions about understanding, and assessment should be a learning opportunity, not just measurement.

Per-competency evidence: Delegation shows in goal-setting, task breakdown, and tool selection. Description shows in conversation quality and iterative refinement. Discernment shows in critical evaluation. Diligence shows in ethical decision-making and accountability.

Designing assignments

Three design principles: authenticity (mirror real-world collaboration, not classroom-only rituals), iteration (refinement is where growth shows), and pedagogical transparency (assess the collaboration process itself, openly, not just outputs).

Assignment types worth stealing

• Improve an AI output: hand students a mediocre AI draft and grade the improvement.
• Compare AI systems on the same task.
• Annotated chat logs: students mark up their own conversations, naming what worked.
• Recorded narrations of the process.
• Learning journals across a term.
• Personal policy statements: students write their own values, strategies, and methods for AI collaboration. These work on every level at once: students develop their own approach, feel empowered and accountable, and reflect deeply.
• Peer review, which develops Discernment through evaluating others' work.

Volume management, because process assessment is heavy: detailed rubrics, peer review, lightning-round conferences, and selective sampling of the full logs.

AI's impact on your discipline

Three questions every discipline must now answer, and all three are worth asking:

1. What gets automated in your field?
2. Where does human-AI partnership add the most value?
3. How do students stay accountable for AI systems in their future careers?

Disruption is not uniform. Some disruptions are opportunities to leverage, others are problems to solve, and your disciplinary expertise is what determines which is which. Build, as a designed exercise, your ready answer to the question students will actually ask: "Why learn this when AI can do it?"

Making tacit knowledge explicit

Experts hold most of their standards tacitly, and AI-era teaching requires surfacing them. With colleagues, work through each D:

• Discernment: what does quality actually look like in our field, beyond vague terms?
• Description: what are the real artifacts of our work, and what are the expert thought processes behind them?
• Delegation: decompose the field's work and decide, element by element: automate, augment, or agent?
• Diligence: codify the field's ethics, disclosure norms, and accountability expectations.