How do I discuss AI benefits and risks in the same meeting?
Put the benefit and the risk in the same proposal for one specific use of AI. Name the task and what a useful result looks like, then name a failure in that same workflow. That keeps the meeting on one use, and away from a debate over all AI.
The Trust Imperative conversation kept returning to opportunity and risk. Put both in the pilot proposal, so nobody has to pick a side.
Say an AI agent drafts internal answers from a policy library. A useful result might be a draft answer with the relevant policy passages. A failure might be an answer built on an older policy. Say who checks the source and what happens when the documents disagree.
Keep your assumptions visible. If you haven't measured how long the task takes now, call any expected time saving a hypothesis. If you haven't tested a boundary, don't describe it as proven protection.
A compact brief makes the discussion concrete. People can support the benefit and still ask hard questions about the design. Give them enough detail to do both before the work expands.
What to try this week
Write a compact pilot brief for one use of AI your team is considering: the task, the expected benefit, the failure to test, the review owner and the condition for stopping the pilot.