AI Literacy Faculty Playbook

Teaching AI Judgment, One Habit at a Time.

BE Critical

Stop. Think. Then decide.

I Can: I can pause before using AI-generated content, evaluate it for accuracy, bias, and credibility, and explain my reasoning.

Building Professional Judgment · Student AI Literacy

Purpose

AI generates output quickly and confidently — but speed and confidence are not the same as accuracy or reliability. BE Critical is the habit of inserting a deliberate pause between receiving AI output and acting on it.

Students who practice BE Critical don't just accept what AI produces; they interrogate it. They ask whether it's accurate, whether it reflects bias, whether it actually answers what was asked — and they can explain their thinking. In an information environment full of plausible-sounding errors, this habit is essential in every discipline.

Students develop this capability through repeated low-stakes practice — evaluating AI output in contexts where the consequences of getting it wrong are manageable, before they face high-stakes decisions.

This is a coaching resource — not a grading rubric unless intentionally connected to a course learning outcome.

Developmental Progression

Introduction

First-Year Experience courses

Students receive AI output and tend to use it. The pause between getting output and acting on it hasn't developed yet — AI responses feel like answers rather than starting points. The goal is introducing the habit of stopping: naming the expectation that AI output requires evaluation before use.

Exploration

First-year and general education courses

Students are developing the habit of pausing before they use AI output. They check some things, question some responses, and can often tell when something doesn't quite fit — even if their evaluation isn't yet consistent or systematic. They're starting to recognize that AI output is a draft, not a verdict.

Application

Program, clinical, shop, and workplace settings

Students stop before using anything AI generates. They check it, question it, and ask whether it actually answers what they needed. They can stand behind what they use — and if they can't, they don't use it. Their evaluation draws on what they know about their field, and it is consistent, systematic, and transferable to new tools and contexts.

Progression is developmental, not linear. Students may move fluidly between stages depending on the task, the tool, or the course context. The settings under each stage are typical examples, not placements.

I Can Statements

I can pause before using AI-generated content and ask whether it is accurate, biased, or incomplete.
I can evaluate AI output for credibility — checking claims, noticing gaps, and identifying where the output might mislead.
I can explain my reasoning for using, modifying, or rejecting AI output.

What This Looks Like · Habits in Practice

Pauses before submitting or acting on AI output — treating it as a starting point, not a final answer
Can identify at least one specific concern about a piece of AI output — not just a general sense that "AI can be wrong"
Checks AI-generated claims against other sources before relying on them
Modifies or rejects AI output when it doesn't meet the standard needed — and can explain why
Asks "does this actually answer my question?" rather than accepting the first response

Where This Fits Naturally

Research assignments where students use AI to gather, summarize, or interpret information
Writing tasks where AI is used for drafting, outlining, or editing
Technical or clinical problem-solving where AI output could influence a real decision
Information literacy activities using the SIFT method or similar evaluation frameworks
Any assignment where students are asked to evaluate, analyze, or synthesize information from multiple sources
Career research and decision-making activities where students assess the quality and fit of AI-generated information

Blackboard Ultra Quick Wins

These are examples — not requirements. Choose the ideas that fit naturally into your course, assignment, and students' level of experience.

Discussion

Post an AI-generated response with a deliberate error or bias and ask students to find it — then discuss how they would have caught it without being told it was there.

Sample Prompt

"Here's an AI response to a question relevant to our course. Something in it is inaccurate or misleading. Find it, explain why it's a problem, and describe how you would have caught it on your own."

Journal

Ask students to use AI to answer a course-relevant question, then document their evaluation process — what they checked, what they trusted, and what they changed or rejected.

Sample Prompt

"Ask an AI tool a question related to our course topic. What did it get right? What did you question? What did you verify — and how? What would you not use, and why?"

Bb Video Studio

Have students record a think-aloud evaluation of an AI response — narrating their reasoning as they work through whether to trust, modify, or reject the output.

Sample Prompt

"Record yourself evaluating an AI response to a course-relevant question. Talk through your thinking: what you're checking, what concerns you, and what you would do with this output."

Assignment

Give students an AI-generated draft or summary relevant to your course and ask them to annotate it — marking what's accurate, what's questionable, and what needs verification or correction.

Sample Prompt

"Here is an AI-generated summary of [course topic]. Annotate it: highlight what you'd keep, flag what needs verification, and mark anything you'd remove or correct. Explain your reasoning for each."

AI Chat Conversation

Have students use AI to answer a question, then deliberately push back — asking follow-up questions, challenging claims, and documenting where the AI held firm, revised, or revealed its limits.

Sample Prompt

"Ask an AI a question relevant to our course. Then challenge at least two of its claims with follow-up questions. Document how it responded — and what that tells you about how much to trust it."

Knowledge Check

Present students with two AI responses to the same question — one accurate, one subtly flawed — and ask them to identify which is which and explain their reasoning.

Sample Prompt

"Here are two AI responses to the same question. One is accurate; one contains a significant error. Which is which? How did you decide? What would you have needed to catch this in the real world?"

Supporting Student Growth

The hardest part of BE Critical is breaking the reflex to accept. Students who have always been rewarded for finding answers quickly need explicit permission to slow down — and explicit coaching on what slowing down actually looks like.
Teach evaluation as a skill, not a suspicion. The goal is not to make students distrust AI — it's to help them develop the judgment to know when to trust it and when not to.
In technical and health sciences courses, the stakes of acting on bad information are high and concrete — use those stakes to make the habit feel necessary, not academic.
The most powerful evaluation question is "how would I check this?" — not "is this right?" Build that question into your feedback and discussion habits.
Students often don't know what to look for. Give them field-specific evaluation criteria — not generic "check your sources" advice, but standards for what accuracy and credibility look like in your discipline.
Low-stakes practice is essential. Students need many opportunities to evaluate AI output in manageable contexts before they face high-stakes decisions.

Student Reflection Prompts

Describe a time you used AI output without questioning it. Looking back, should you have paused? What would you have checked?
What does it feel like to evaluate AI output critically — versus just accepting it? What gets in the way of the pause?
What are the specific things you look for when you evaluate AI output in your field? How did you learn to look for those things?
Has AI ever given you something that sounded right but turned out to be wrong or misleading? What did you learn from that?
If you couldn't use AI for a task, what would you do instead? What does that tell you about what AI is actually contributing?