Prompts and directions for instructors facilitating group discussion
Overview
This module is about a concept — trust — that is deeply personal and relational. The discussion should create space for people to think out loud about their actual professional relationships, not abstractions. The goal is to help students move from intellectual understanding to concrete insight about their own situation.
Some students will be defensive about this material. They will argue that accountability is not really possible, that AI systems are not really unreliable, that trust is not really necessary. Listen to that resistance with respect. But also challenge it. The point of this module is to see clearly, not to hide.
Pre-Module (Before Lesson 1)
Opening Activity: What Is Trust?
Time: 15 minutes Format: Large group or small groups
Ask: "Without looking at the course material, define trust. What does it mean to say you trust someone professionally?"
Capture the definitions. Look for patterns. Do people emphasize consistency? Reliability? Predictability? Do they mention values, character, judgment? Do they talk about outcomes or relationships?
Why this matters: This establishes what students bring to the conversation before they learn the module's framework. It also surfaces the difference between reliability (which can be technical) and trust (which is relational). Use this list to return to in final discussion.
Lesson 1: The Three Kinds of Reliability
Discussion Prompt 1: Where Is AI Already in Your Work?
Time: 20 minutes Format: Small groups or pairs, then large group
Ask students to name a tool or system they already use in their work that involves AI or automation (even if they would not necessarily call it "AI"). It could be software that flags errors, a template system, a search tool, a scheduling algorithm, anything.
Then ask: "Is this tool reliable? In what ways — consistency, accuracy, availability? Where does it fall short?"
What to listen for: - Do students distinguish between these kinds of reliability? - Do they assume that consistency means it is trustworthy? - Can they name gaps where the tool fails in judgment or context?
Follow-up: "If this tool made a decision that affected your credibility, would you be comfortable having that decision attributed to the tool, or would you need to own it?"
Discussion Prompt 2: The Limits of Consistency
Time: 15 minutes Format: Large group discussion
Pose this scenario: "A document summarization tool consistently produces summaries that are grammatically correct, well-structured, and accurate to the source material. But the summaries always emphasize technical details and miss the strategic implications. Is this tool reliable? Can you trust it with client-facing work?"
What to listen for: - Do students see that consistency is not the same as usefulness? - Can they articulate what is missing (judgment about what matters)? - Do they see that someone still has to validate the output?
Lesson 2: What Trust Actually Requires
Discussion Prompt 3: Stakes and Accountability
Time: 25 minutes Format: Small groups (3–4 people), then report back
Divide students into groups. Give each group one of the four prerequisites of trust: 1. Shared stakes 2. Accountability 3. Context 4. Demonstrated history
Ask each group: "In your own work, what does this look like? Give me an example of a relationship where this prerequisite is strong, and an example where it is weak."
What to listen for: - Can students give concrete examples? - Do they see how these prerequisites are interconnected? - Can they describe what would change if one of these were missing?
Debrief with large group: As groups report, draw a simple relationship map on the board showing how the four work together. Emphasize that all four have to be present, not just one or two.
Discussion Prompt 4: Why AI Fails on These Fronts
Time: 20 minutes Format: Large group discussion
Ask: "For each of the four prerequisites, explain why an AI system structurally cannot meet it. Not 'current AI' but 'any AI system.'"
Push back if students say "but maybe in the future" or "but with the right design." The point is to reach clarity that this is not a technical limitation but a structural one. An AI system, by definition, cannot hold stakes, be held accountable, understand context in the full sense, or have demonstrated character.
What to listen for: - Do students grasp the structural nature of the limitation? - Are they comfortable with the idea that some things cannot be automated? - Do they see why this creates new kinds of professional demand?
Lesson 3: The New Trust Demand
Discussion Prompt 5: Who Is Accountable?
Time: 20 minutes Format: Small groups, then large group
Pose scenarios and ask: "In this situation, who should be accountable?"
Scenarios: 1. "An AI-powered hiring tool screens job applications. It flags candidates with high likelihood of staying with the company. A hiring manager uses this tool and makes decisions based on it. A candidate who was screened out claims discrimination. Who is accountable?"
-
"A customer service AI is deployed to answer routine questions. It gives an answer that is technically accurate but misleading in context. A customer makes a business decision based on that answer and loses money. Who is accountable?"
-
"A team uses AI to draft project proposals. The proposals are well-written and persuasive. A manager sends them to stakeholders with minimal review. The proposals promise something that turns out to be impossible. Who is accountable?"
After small groups discuss, gather large group and ask: "What do all three scenarios have in common? What does accountability look like in an AI-assisted world?"
What to listen for: - Do students default to blaming the AI? - Can they articulate why human accountability is still necessary? - Can they see where accountability got fuzzy and what to do about it?
Discussion Prompt 6: The New Role
Time: 15 minutes Format: Large group discussion
Ask: "In your organization, who holds the role of deciding where AI is used? What is their title? Are they a technical person? An executive? Multiple people? Is it clear?"
Then ask: "What do they need in order to do that job well? Not what tools, but what knowledge, what relationships, what authority?"
What to listen for: - Do students see this as a real and growing role? - Can they articulate what qualities that role requires? - Do they see this role as a constraint or as an opportunity?
Lesson 4: Accountability as a Professional Asset
Discussion Prompt 7: Owning vs. Obscuring
Time: 20 minutes Format: Small groups with role-play element, then discussion
Put students in pairs. Give them this scenario:
"You use an AI tool to draft a report. The report is good, but there are three sentences where you are not confident in the accuracy. You catch one error and fix it. For the other two, you are not sure. You have two options:
A) Send the report with no mention of the AI. If the two uncertain sentences turn out to be fine, you look great and the report will be seen as entirely your work. If they turn out to be wrong, people will find out eventually that you used AI and did not fully check it.
B) Send the report with a note explaining that it was drafted with AI assistance, that you caught and fixed one error, and that you have these two remaining uncertainties about specific sentences.
Which do you choose? Why?"
Have pairs discuss. Then ask them to predict: "Which choice will build more trust over time?"
What to listen for: - Do students gravitate toward the short-term benefit or the long-term relationship? - Can they articulate why transparency, even about uncertainty, builds trust? - Do they see this as a sustainable approach?
Follow-up: "Have you ever made this choice explicitly? What happened?"
Discussion Prompt 8: The Cost of Blame-Shifting
Time: 15 minutes Format: Large group discussion, story-based
Ask students: "Tell me about a time when someone blamed a tool or a system for a failure that was actually their responsibility for how they used it. How did that affect your trust in them?"
Listen to stories. Do not interrupt. Then ask: "What if they had owned it instead? Would you trust them more?"
What to listen for: - Do students see the pattern of blame-shifting eroding trust? - Can they articulate what changes when someone accepts responsibility? - Do they see accountability as strength or weakness?
Lesson 5: Building Trust Deliberately
Discussion Prompt 9: What Does Transparency Look Like?
Time: 20 minutes Format: Small groups, then large group
Ask: "You just produced a piece of work using AI assistance. You need to send it to [your boss / your client / your team]. What do you tell them about how it was produced? Write out what you would actually say."
Have groups draft language. Share across groups. Discuss which versions build trust and which deflect or obscure.
What to listen for: - Do students default to hiding the AI, or being transparent about it? - Can they find language that is honest without being self-deprecating? - Can they articulate the difference between transparency and over-explanation?
Discussion Prompt 10: Consistency as a Professional Practice
Time: 20 minutes Format: Large group discussion
Ask: "What does it mean to be consistent about how you handle trust? Give me an example of a consistent practice that builds trust."
Then ask: "What are your current inconsistencies? Where do you do something one way with one person and a different way with another? Is that by design or by accident?"
What to listen for: - Do students see consistency as a deliberate practice? - Can they articulate the tradeoffs (consistency can be boring but it is reliable)? - Can they identify one area where they want to be more consistent?
Discussion Prompt 11: Relationship Investment
Time: 15 minutes Format: Pair discussion, then large group
Ask pairs: "Who is one person in your professional life who you need to build more trust with? What would change if you invested more time in that relationship? What would you have to give up?"
After pairs talk, ask for volunteers to share. Do not force it. But listen to the patterns. Do people see relationship investment as necessary or optional? Do they see it as something that can be automated or something that requires genuine presence?
Post-Module (After Exercises and Deliverable)
Final Discussion: Your Trust Inventory
Time: 45 minutes Format: Small groups, then large group sharing (optional)
Ask students to bring their Trust Inventory (or at least the thinking from it). Do not ask them to share specific content if they are uncomfortable. But ask them to share patterns.
Large group prompts: - "What did you learn about yourself from doing the Trust Map?" - "Were there gaps between how you think you handle accountability and how you actually do?" - "What surprised you about your trust inventory?" - "What is one behavior you are committing to change?"
What to listen for: - Do students see themselves clearly, or are they still defending? - Are they specific about what they are going to do, or vague? - Do they see trust as something that develops or something they either have or do not have?
Closing: Connecting Back to "One Skill to End Them All"
Ask students to re-read the passage from Michael Bilca's essay about trust:
"AI can execute tasks. It cannot hold accountability. It cannot be promoted. It cannot be trusted with a budget, a team, a crisis, a client relationship — not in the full sense of the word trusted, where trust means 'I believe this entity will do the right thing even when no one is watching and even when the right thing is hard.'"
Then ask: "How does that passage land for you now? Does it make sense? Do you see it in your own work?"
What you are doing: Closing the loop. Showing students that this module is not abstract philosophy — it is a description of what actually happens in organizations.
Facilitation Notes
For the defensive student: Some students will argue that this is all too pessimistic about AI, or that accountability is not really possible in modern organizations. Listen with respect, but also be clear: the module is not arguing against AI. It is arguing for clarity about what AI is for and what humans are for. Accountability may be difficult in modern organizations, but the module argues it is increasingly valuable because it is rare.
For the student who wants to move fast: Some students will want to rush past the relational, messy stuff to abstract principles. Slow them down. Push them to name relationships, specific outcomes, actual choices. The value is in the specificity.
For the student who is struggling: This module can hit hard if you are in a period of your career where trust is fragile or relationships are difficult. Create space for that. You do not have to share. But acknowledge that trust is real and matters.
For the whole group: Emphasize repeatedly that this module is not about judgment. It is about seeing clearly. Everyone has relationships where trust is strong and areas where they need to build it. That is normal. The point is to see it and do something about it.
Discussion Timing
If you have limited time: - Prioritize Discussion Prompts 1, 3, 5, 7, and 11 (the most relational and practical) - Condense others or assign as reflection rather than group discussion
If you have extended time (multiple sessions): - Add breakout room discussions where students talk one-on-one with someone they trust about trust - Have students draft their Trust Inventory in session and get feedback from peers - Invite a guest (someone who has built significant trust in an AI-assisted role) to share their story
Resources for Instructors
- Michael Bilca, "One Skill to End Them All" (provided in course materials) — the conceptual foundation
- Examples of organizations making accountability clear around AI use (from compliance, ethics, or AI governance literature)
- Case studies of failures where accountability was unclear (provides real-world grounding)