Human-Centered Collaboration

Group Dynamics

With an AI-intensive activity, how do we ensure everyone engages with the AI rather than one person becoming the "AI operator"?

Propose that as the groups move through the "hats," within each group, assign rotating roles:

The Prompter — interacts (types/speaks) with AI
The Challenger — questions assumptions. AI tools will often silently make assumptions when they lack context. Keeping this at the front of mind helps recognize this when it happens.
The Synthesizer — captures human insights. Put the "human" in human-centered AI.
The Keeper — maintains big-picture vision. It can be easy to go down "rabbit-trails" and get into the weeds.

For each hat, explicitly require human work BEFORE AI use. For example:

Designer Hat

5 minutes of team discussion/observation sharing before any AI interaction

Builder Hat

Sketch rough ideas on paper before asking AI to expand

Prototype Phase

Create low-fi prototype elements before AI enhancement

Advocate Hat

Outline key points before AI helps structure

Add micro-reflections between prompts:

"What did the AI add/change/skip that surprised you?"

"What did you know that the AI missed?"

"How would you modify this prompt next time?"

Watch out for these warning signs:

Designer Hat

AI generates problem statement → team can't explain why it matters

Builder Hat

AI creates perfect journey map → team hasn't built empathy

Prototype Phase

AI builds complete prototype → team can't iterate on it

Advocate Hat

AI writes pitch → team delivers without conviction