AI Usage Policy Development Kit

As of this writing (2026). Institutional and course-level AI policies are still evolving — revisit this page periodically rather than treating it as permanent guidance.

Definition

Rather than following someone else’s blanket AI policy, this is a habit of writing your own: answering five specific questions to produce an explicit, personal policy for how you’ll use AI tools in your work. This matters more, not less, as institutional policy stays vague — most schools and workplaces don’t have a specific policy for generative AI; use typically falls under general honesty and misconduct rules instead, which tell you what’s prohibited without telling you how to use permitted AI assistance well. Writing your own policy is what closes that gap: a forcing function for thinking carefully about how you use AI and how you make sure the result is still genuinely yours, connected to the same authorship principles covered in About Corpus → AI as Collaborator.

Learning Outcome

After using this technique, you should have an explicit, written personal policy — not just a vague sense of “I try to be honest about it” — that you could actually hand to someone else and follow yourself under pressure.

Core Structure

Five questions, each demanding a specific, concrete answer rather than a general intention:

1. Transparency. How will you ensure transparency when using AI in your work? (Not “I’ll be honest” — a specific practice: what gets disclosed, where, in what format.)
2. Boundaries. In what situations will you avoid using AI tools entirely? (Not “when it feels wrong” — specific situations you can recognize in the moment.)
3. Verification. How will you verify the accuracy and fairness of AI-generated output? (A specific check you’ll actually perform, not just “I’ll double-check it.”)
4. Privacy. What privacy and data protection practices will you follow when using AI? (What you will and won’t paste into a tool, and why.)
5. Disclosure. How will you disclose AI involvement in your deliverables? (The actual language or format you’ll use, not just “I’ll mention it.”)

Worked Example

A sample answer to question 1 (transparency), specific enough to actually follow: “I’ll keep a running note of every AI-assisted step in a document alongside my draft — what I asked, what I got back, and what I changed before it went into the final version — so I could reconstruct my process for anyone who asked, not just summarize it from memory afterward.” That’s concrete enough to actually do, unlike “I’ll be transparent about it,” which isn’t a practice at all.

Common Pitfalls

• Answering all five questions with some version of “I’ll be honest / careful / responsible” — none of that is a policy, it’s a value statement with no practice attached.
• Setting boundaries so vague they’d never actually stop you in the moment (“I won’t overuse it”).
• Treating verification as a one-time check rather than a repeatable practice you’d apply to every piece of AI-assisted work.
• Writing the policy once and never revisiting it as tools, coursework, or your own judgment change.

Checklist

• Transparency practice is a specific, repeatable action, not a value statement
• At least one concrete situation is named where AI won’t be used at all
• Verification has an actual method attached, not just an intention
• Privacy practice specifies what will and won’t be shared with a tool
• Disclosure has a real format or phrasing, ready to actually use