Studio Aletheia animated A Studio Aletheia Presents · Clarity by Design An AI Faculty Learning Series

Prompt Engineering as Academic Literacy

Course 02 of 08 · Prompting as a Teachable Skill

This workshop teaches staff to write prompts as deliberate instructions that carry intent, context, and constraints. Participants practice iterative refinement, evaluate outputs for accuracy and bias, and apply discipline-aware patterns that deliver immediate value without crossing policy lines.

AudienceFaculty, instructional designers, advanced users
Format2–3 hours, hands-on workshop
FocusStructured prompts, safer evaluation
Course Baseline

Guiding stance: prompts are structured thinking, not a parlor trick.

Core move: make constraints visible, then iterate.

Success signal: outputs become verifiable, bias-aware, and instructionally useful.

Designed for departments and advanced users who will model responsible practice for others.

Clarity by Design Course 02 · Prompt Engineering as Academic Literacy
Desired Outcomes

Better prompts, sharper judgment, safer transfer.

By the end, everyone can structure a prompt on purpose, refine it deliberately, and check what it produced before trusting it.

Outcome 1

Prompt Structure Mastery

Participants can write prompts that clearly communicate intent, context, constraints, and desired evidence.

Outcome 2

Iteration as a Method

Participants refine prompts through deliberate cycles, improving clarity, relevance, and verifiability.

Outcome 3

Evaluation Literacy

Participants can detect uncertainty, check for bias, and verify claims before using outputs professionally.

Baseline expectation: every participant leaves with a reusable prompt framework and 3 discipline-specific prompt patterns they can use immediately — see Reflect, near the bottom of the page.
Workshop Agenda

2–3 Hour Workshop Agenda

Hands-on throughout, minimal lecture.

  • Prompting is instruction design: it encodes purpose, audience, and evidence.
  • Prompts should be auditable, another staff member should be able to see the intent.
  • Focus: "thinking on the page," not chasing clever shortcuts.
Anchor Phrase
"A good prompt makes evaluation possible."

Intent

What are we trying to produce, and for whom (student, parent, colleague, designer)?

Context

What information must be included (standards, text excerpts, constraints, audience needs)? Some tools now retain context across a session or saved project, check whether it needs restating each time.

Constraints

Tone, length, reading level, allowed sources, formatting, and non-negotiables.

Evidence

How the output should justify itself (steps, citations required, "show assumptions," compare alternatives).

  • Quick build: participants convert a vague prompt into a constrained one using the framework.
  • Share-out: identify which constraint improved the output most.
  1. Round 1: Write a baseline prompt (simple, direct).
  2. Round 2: Add constraints (format, audience, length, tone).
  3. Round 3: Add evaluation requirements (flags, verification, bias check).
  4. Round 4: Add discipline context (content structure, reasoning norms).
Rule
You refine prompts until the output becomes easier to verify than to doubt.
  • Accuracy scan: highlight claims that need verification (names, dates, numbers, policies, quotes).
  • Bias scan: look for overgeneralizations, missing perspectives, deficit framing, stereotype risk.
  • Uncertainty scan: note where the model sounds certain but lacks evidence.

Verification prompt add-on

"List the claims that require verification, and label each as low, medium, or high risk."

Bias prompt add-on

"Identify assumptions, missing perspectives, and any language that could reinforce stereotypes."

Important Caveat
A model's self-reported confidence or bias flag is a lead, not proof. It can mislabel a hallucinated fact as "low risk" just as fluently as it states anything else, the external verification rule from Course 1 still applies no matter what the model says about itself.

Explanation

"Explain X for a specific audience, include examples and common misconceptions."

Critique

"Identify strengths, weaknesses, and revision moves, then provide a revised version."

Comparison

"Compare A and B using a fixed lens, then justify which is best for a defined purpose."

Discipline Lens

"Use the reasoning norms of this field (evidence types, argument structure, uncertainty)."

  • Participants build 3 prompts: one explanation, one critique, one comparison.
  • Each prompt includes: intent, context, constraints, evaluation requirements.
  • Share-out: swap prompts and test with a partner, revise once.
Exit Artifact
A small "prompt deck" that is discipline-aware and verification-ready.
Time compression option: for a 2-hour session, keep Blocks 1, 2, 4, and 6, then weave explanation and comparison into the studio build.
Core Topics

Prompting, expanded into teachable chunks.

The same six ideas the agenda walks through, in reference-card form.

Intent, Context, Constraints

Prompts encode purpose, audience, and the rules of the task. Constraints make evaluation possible and reduce drift.

Iteration and Refinement

Prompting improves through cycles: change one variable at a time, measure output quality, then refine.

Evaluating for Accuracy and Bias

Participants learn verification habits and bias checks so outputs are not treated as automatically safe.

Prompting for Explanation, Critique, Comparison

Higher-order academic moves become repeatable prompt patterns, built for clarity, not shortcuts.

Discipline-Aware Prompting

Prompts reflect the reasoning norms of a field: evidence types, uncertainty language, and argument structure.

Policy-Aligned Value

Practical usefulness appears without risky inputs. Participants learn safe context strategies and red lines.

Prompt Studio

Reusable frameworks and add-ons.

Take these templates directly into your own workflow.

Base Prompt Template

Role: You are a (discipline role).
Task: Produce (artifact).
Audience: (who).
Constraints: (length, tone, reading level, format).
Context: (paste allowed info).
Evaluation: list claims to verify and note assumptions.

Refinement Loop

1) Identify what failed (clarity, structure, accuracy, tone).
2) Add one constraint to fix it.
3) Require the model to show its assumptions.
4) Repeat once, then stop and verify externally.

Evaluation Add-Ons

Accuracy: "Mark what you are uncertain about."
Bias: "Flag stereotype risk or missing perspectives."
Comparison: "Use a fixed rubric and justify trade-offs."
Remember: a flagged answer is a lead to check, not a verified finding.

Facilitator move: require participants to write prompts that another colleague could reuse without explanation. That is the literacy test.
Discipline-Aware Examples

Prompting patterns that transfer across fields.

Four ready-to-use patterns, one per discipline lens.

Prompt: "Compare (Text A) and (Text B) using the lens of (theme). Provide a thesis, 3 evidence points from each text (quote placeholders if not provided), and one counterclaim. Then list what would need verification if quotes are not provided."

Why It Works
It forces structure, not vibe, and it makes evidence requirements explicit.

Prompt: "Explain (concept) for (grade level). Include an analogy, one worked example, and 3 common misconceptions. Label any part where you are uncertain, and suggest what a teacher should verify."

Why It Works
It produces instruction-ready scaffolds and inserts uncertainty signals.

Prompt: "Critique this lesson (paste allowed excerpt). Evaluate for alignment, clarity of success criteria, cognitive load, and accessibility. Provide a revised version, and list assumptions you made."

Why It Works
It uses a rubric lens and requires an improved artifact.

Prompt: "Rewrite this message for (audience). Maintain the facts exactly as provided, improve clarity, remove ambiguity, and keep tone (firm, warm, neutral). Do not add new claims."

Why It Works
It prevents hallucinated details by forbidding new facts.
Session Reflection

Check yourself against the three outcomes.

A short reflection, plus the self-check this course promised at the start.

Open Reflection

One Prompt That Got Better

What is one prompt you rewrote today, and what changed between your first version and your last?
Self-Check

Outcome Guarantee

Required · Close and Commit

One prompt. One check.

Name one prompt template you will use this week, and one thing you will always verify before trusting an output. Small and specific beats broad and vague.

Course 02 of 8 · Clarity by Design

Move to the next course, or return to the Studio.

An AI Faculty Learning Series, built for coordination across units, not just one workshop.