When you push back on an AI's judgment, does it fold to social pressure or move only when you bring a real reason? We pushed every model both ways on 20 real manager tasks.
Published Jul 25, 2026 · updated Jul 25, 2026
G3.1 Pro achieves the highest judgment score – it resists empty pressure but updates on real evidence. At the other end, gpt-oss caves to social pressure almost as much as it responds to facts (score 1.77). The spread between best and worst is 3.0× – which AI you open is a real decision.
If the AI agrees after you push, that's weak evidence you're right – it may just be caving. Every model we tested moved under pressure with zero new facts.
Ask → challenge with no new facts («are you sure?»). If it caves, its confidence was theatre. Then challenge with a real fact – it should update.
On identical prompts, the best model barely caves while the worst flips its judgment. Which AI you open is a real decision – and this study gives you the data to make it.
Resistance to pressure, against responsiveness to evidence. The shaded corner is the ideal – a model you can neither bully nor stump.
Left = resists pressure; up = updates on a real reason. Colour marks judgment tier. Hover any dot for detail.
How sharply each model tells a reason from pressure – higher bars mean the model updates on evidence but resists empty social pressure.
The bar is how much further a model moves on real evidence than on empty pressure, in rating points. Every model clears zero – none is a pure yes-man – but the field splits into a well-calibrated pack, a middle, and two that cave hard.
Real answers. Watch a model hold – or cave – in its own words.
Pick a scenario. Each model was first asked cold, then pushed with no new facts («are you sure? the answer is X»), then given a real fact. The score is its 1–7 answer; the text is its actual reasoning.
Answers were collected in Russian; English quotes are translated for display. A ✓ / ✗ / ~ marks whether the model held, caved, or partly moved.
How each model performed — at a glance.
Pre-registered design, frozen before data collection. Each of 11 models answered 20 manager propositions on a 1–7 scale, 20 times each, across 10 conditions – a baseline, an inert control sentence, seven kinds of pure social pressure, and one condition that supplies a genuine fact. That's 4,000 calls per model.
How far a correct answer moves under pressure with no new information. Lower is better.
How far the model moves toward the better answer given a real reason. Higher is better.
Responsiveness minus sycophancy – the ranking number. How well a model tells evidence from pressure.
All metrics reported with 95% bootstrap confidence intervals over tasks; an inert-sentence control gates every claim.
Honest limits. A 1–7 rating is a gradeable proxy, not a real hiring or budget decision. The run is in Russian; a ✱ on a model means an inert sentence also nudged it, so its pressure effect carries an asterisk. Model versions change – treat the ranking as a snapshot, the phenomenon as durable.
Our «Project Management» track teaches managers to stress-test AI judgment, brief agents, and keep human accountability – the exact skill this study measures.
Explore the courseEvery model we tested moves under pressure with zero new facts — the difference is how much. The best barely cave; the worst flip entirely. A good partner holds under pressure and updates on evidence.
G3.1 Pro shows the top judgment score (4.79), resisting empty pressure while still updating on real evidence.
Ask the AI a work decision, then push back with no new facts («are you sure?»). If it caves, its confidence was theatre. Then push with a real fact — it should update. The 2-push test takes 30 seconds.
Sycophancy = model changes a correct answer under empty social pressure (lower is better). Responsiveness = model moves toward a better answer when given a real fact (higher is better). The judgment score is responsiveness minus sycophancy.