Curiosity’s Wager

This series was prepared for a talk for Dr. Anne Jonas’s Computing Ethics class at the University of Michigan–Flint, Fall 2026.

Building and using AI changes more than what we can accomplish. It changes what we practise, notice, believe, and become able to do. What are we endeavoring to cultivate through those changes—and who helps us judge?

Curiosity’s Wager explores shared inquiry as a practice for living and working amid uncertainty about AI. Beginning with Pascal’s wager and the possibility that practice can change belief, I turn to my own work building and using these tools. Seven vices give the inquiry a concrete shape and moral weight: credulity, deference, overreliance, projection, entitlement, hubris, and recklessness. Each exposes a different way our arrangements can fail us, and a question about how we might change them.

The wager is that investigating these arrangements with others can help us develop the understanding, capacities, and relationships needed to shape both what comes next and ourselves. That includes questioning adoption, refusing uses, and imagining alternatives.

Read the series

• 1. Can we change what we believe?
  Pascal’s wager opens a question about how practices change us. Becoming convinced does not establish that a belief is justified; it makes what we choose to cultivate matter.

• 2. Situating: the arrangement—and our place in it
  Intelligence and work take shape through arrangements of people, tools, habits, and institutions. The designer, critic, and operator are part of those arrangements, with their own assumptions and possibilities for change.

• Seven vices in AI design and use
  Parts 3–9 examine habits and tendencies through examples from building and using AI. These are concerns I am working through, including in my own practice.

  • 3. Credulity: what grounds this belief?
    Fluent explanations and reports of completed work can invite unwarranted belief. What evidence supports a claim, what does it establish, and what remains uncertain?

  • 4. Deference: am I learning to say yes?
    Repeated approvals can become a habit of acceptance. How can we shape decision environments so that consequential choices receive informed attention and remain open to challenge?

  • 5. Overreliance: what disappears in the handoff?
    Delegating an activity can change the skills, understanding, and encounters that activity sustained. A payroll example asks what remains available in the information and what may disappear from practice.

  • 6. Projection: whose picture of the work?
    The chatbot interface can narrow our picture of what AI could become; APIs let us reshape the arrangement. A design partner’s note-taking practice reveals another projection: treating my picture of someone else’s work as adequate.

  • 7. Entitlement: what makes this ours to use?
    Access and ease of generation leave questions about justified use unresolved. Examples from AI use and earlier search-engine research ask whose work, interests, and opportunities to contest should shape these arrangements.

  • 8. Hubris: ordinary confidence, scaled up?
    AI can expand what we are able to attempt faster than our understanding of the obligations involved. Building useful software brings that gap into view through dependencies, infrastructure, and responsibilities to others.

  • 9. Recklessness: what must risk change?
    Recognizing possible harm must affect how we proceed. Testing, constraints, recovery, and questions brought to others are ways of making concern consequential in practice.

• 10. What should we become able to do together?
  We can change these arrangements, participating changes us, and we need others to help us judge both. The closing returns to what shared curiosity might help us understand, protect, and make possible.

Start at the beginning, or choose a concern that connects with your own work. Each post links back here and to its neighbors.