Example output
# Focus Group Moderator Guide
## Topic: Perceptions of Generative AI Tools Among Early-Career Journalists
## Participants: 8 journalists, 1-5 years experience, mixed print/digital
## Duration: 90 minutes
---
### 1. Welcome and Ground Rules (5 min)
*Moderator stands at whiteboard, warm but purposeful tone.*
"Thanks so much for coming tonight. I'm Nora, and I'll be your moderator. The goal of the next 90 minutes is pretty simple — I want to understand how journalists in the first few years of their career are thinking about, using, or avoiding generative AI tools in their work. There are no right or wrong answers. I am not from a company that makes these tools, and nothing you say tonight will be attributed to you by name.
A few ground rules:
- There's no consensus required. If you disagree with someone, please say so — disagreement is where the interesting stuff lives.
- One voice at a time for the transcriber. If you're quiet, I'll probably gently ask for your view; if you're talkative, I may gently move us on.
- Anything shared in this room stays in this room among us. The recording will be transcribed and de-identified within two weeks.
Let's do a quick round — your first name and what kind of reporting you do."
### 2. Icebreaker (5 min)
"Tell us one tool — any tool, not necessarily AI — that you didn't have a year ago but now couldn't imagine working without."
*This softens the group and establishes a baseline of tool-adoption thinking without priming the AI topic.*
### 3. Main Discussion
**Section A — Current Use (20 min)**
Opening: "Walk me through the last time you used a generative AI tool for work. What were you trying to do?"
Probes:
- "What made that moment feel like the right time to reach for the tool?"
- "Did you tell anyone — an editor, a colleague — that you'd used it?"
- "What would have happened if you hadn't had access?"
**Section B — Trust and Verification (20 min)**
Opening: "How do you decide whether to trust something an AI tool produces?"
Probes:
- "Has there been a moment where you almost published something that turned out to be wrong? What happened?"
- "How is this different from how you evaluate a human source?"
- "If an AI tool got something right 95% of the time, would you use it? 99%?"
**Section C — Professional Identity (15 min)**
Opening: "Some people say these tools are deskilling the profession, others say they're freeing us to do harder work. Where do you land?"
Probes:
- "What does a journalist do that an AI tool cannot?"
- "Has your sense of your own craft shifted in the last 18 months?"
- "What would you tell a j-school student today?"
### 4. Projective Exercise (10 min)
"I'm going to hand you each a sticky note. Draw or write a scene from five years in the future showing you at work. Don't overthink it — two minutes. Then we'll put them up and look together."
*The sticky-note wall often surfaces anxieties and aspirations that direct questioning misses. Look for: presence/absence of AI in the scene, who holds agency, whether newsrooms appear at all.*
### 5. Closing Synthesis (5 min)
"If you were writing the guidance document your newsroom should have but doesn't — three bullet points only — what would be on it?"
### 6. Moderator Debrief (immediately after)
- Who dominated; who deferred?
- What did people not say that you expected them to?
- Which moment had the most energy in the room?
- Any non-verbal reactions that contradicted verbal responses?
- What surprised me?
### 7. Dynamics to Watch
- **Social desirability:** Participants may under-report AI use due to professional norms. Reframe Section A with "tools your colleagues use" as a deflection vehicle if the group is guarded.
- **Dominant voice risk:** If one participant frames AI as purely bad/good early, subsequent speakers will anchor. Intervene with "Let's hear from someone with a different angle."
- **My bias:** I'm sympathetic to craft-preservation arguments. Check my own "mm-hm" affirmations against the transcript.