A practical guide to running an Appreciative Inquiry process focused on ethical AI:

Step 1: Define the affirmative topic
Appreciative Inquiry begins with a topic framed in the affirmative. Not “how do we reduce AI bias?” but “ethical AI excellence in our organisation.” The topic sets the direction of inquiry.

Step 2: Craft appreciative questions
Develop 3–5 questions that invite people to share stories of ethical AI at its best. Test them with a small group before using them more broadly. The best questions feel genuine, not forced.

Step 3: Conduct interviews
Appreciative Inquiry works best through direct conversation: ideally one-to-one or in small groups. Allow people to tell stories in their own words. Listen for values, conditions, and specific behaviours that appear across multiple interviews.

Step 4: Identify themes
Analyse the stories for recurring themes. What values come up consistently? What conditions appear repeatedly in accounts of ethical AI in practice? These themes form the foundation of the Dream phase.

Step 5: Host a Dream session
Bring people together to share what they discovered and to articulate a shared vision of ethical AI excellence. What does it look like when everything you discovered is consistently present?

Step 6: Co-design
Use the dream as a foundation for designing concrete changes to structures, processes, and practices. Ensure that the people most affected by these changes are involved in designing them.

Step 7: Sustain
Build Appreciative Inquiry into ongoing practice. Regular appreciative conversations about ethical AI: not just at crisis points: cultivate the reflective habit that characterises the most ethically mature AI organisations.