StewAI Blog · Decision Science · 16 min read
The Devil's Advocate Problem: Why Assigned Dissent Backfires
Everyone in the meeting nodded, and the plan was a disaster. Four decades of decision science explain why, and why appointing a devil's advocate can make it worse. Then we hand one clean-looking proposal to a StewAI recipe and watch all twelve of its assumptions fall apart.
Frequently asked questions
Why does appointing a devil's advocate often backfire?
Research by Nemeth and others found assigned dissent is discounted precisely because everyone knows it is a role: the group inoculates itself against the objections and walks away more confident, not less. Authentic dissent improves decisions; theatrical dissent protects the plan.
What does the M001 Committee Dissent recipe do instead?
It extracts a proposal's load-bearing assumptions, gathers outside evidence for and against each one, writes an independent critique per assumption in isolation, and assembles a decision brief with per-assumption verdicts. The dissent is structured and evidence-grounded, and no person in the room has to own it.
What happened when a real proposal went through it?
A clean-looking proposal yielded twelve load-bearing assumptions, and the evidence-grounded critique left none of them standing untouched: each came back with specific external evidence and a verdict, turning a nodding meeting into a concrete list of what must be true.
Read the full article on StewAI