False positives are confident
Any single model has blind spots and can state mistakes with the same confidence as facts - there is no signal telling you which is which.
Collaborative intelligence
Ask once. Dr Moot convenes a panel of leading AI models to challenge the answer and each other, expose what's missing and deliver a considered verdict.
The deliberation
Every moot runs in the open. Seats draft, challenge and verify in real time, and the transcript keeps every claim - contested or settled - on the record.
Generator, Sceptic and Specialist work in parallel, each badged with the lab and model holding the seat.
Contested claims stay in the transcript with the vote that split the panel.
The Chair’s ruling links back to the rounds, votes and sources that produced it.
From the dictionary
moot /muːt/ Adjektiv 1. Fraglich, umstritten. Zur Diskussion gestellt. 2. Ohne praktische Bedeutung; rein akademisch. Verb 1. Zur Diskussion stellen; aufwerfen. 2. Erörtern. Substantiv 1. Eine Versammlung des Volkes, einberufen, um zu beraten und zu entscheiden; ein Rat oder eine Zusammenkunft. (Altenglisch mōt, gemōt)
The premise
A single model states its mistakes with confidence. A panel catches false positives before they reach you.
Any single model has blind spots and can state mistakes with the same confidence as facts - there is no signal telling you which is which.
A sceptic challenges the draft, independent labs cross-check it, and only the answer that survives scrutiny is presented - a second opinion by default.
Two claims withdrawn under challenge before the verdict.
Each seat is held by a different lab, so the models genuinely check each other.
The transcript keeps what was claimed, contested and decided on the record.
Auto spots consequential questions and routes them to deeper deliberation.
A dedicated sceptic attacks the draft’s assumptions instead of politely agreeing with it.
Models from separate labs cross-examine each other’s claims, so shared blind spots surface.
Search agents check surviving claims against real evidence before the Chair rules.
Beratung mit Spitzenmodellen von
The panel
Change any prompt. Assign any model. Save the panel and convene it again.

Commits to the strongest complete answer and states its assumptions so the panel can attack them.

Finds factual errors, unsupported assumptions and the logical gaps that could change the answer.

Checks facts, figures, definitions and sources, separating evidence from inference.

Weighs the strongest arguments left standing and delivers one clear, actionable verdict.
How it works
Watch each claim get drafted, contested and refined. The transcript shows the working exchange, while dissent remains visible.
Give the panel your question, context or document once.
Models take defined roles and challenge each other in real time.
Thinking · 3 seats
Cross-review round 2
SUMMARY
Recommendation Do not ban social media for all UK under-16s. The governmen…
The Sceptic and Specialist reinforce my position: causal evidence for population-level harm reduction from a blanket ban is insufficient, while evasion, displacement, privacy costs, and lost benefits are material.
Cross-review round 2
SUMMARY
Where the panel agrees (and why it holds) All three of us independently reject …
Both Generator and Specialist converged on “no blanket ban, regulate design/features” - matching my prior position. No one defended a full ban, so the panel has reached consensus.
Cross-review round 2
SUMMARY
Established evidence Large correlational datasets (Millennium Cohort, YRBS,…
Generator correctly prioritises feature-level duties and privacy-preserving assurance over prohibition; their displacement claim matches my enforcement data.
Claims are checked against sources before the Chair reaches a verdict.
Verification Round 1
CompleteChecked 3 contested claims:
Used 9 sources
Why are France’s top earners leaving and where are they going? | Euronews
euronews.com
Before California, France tried a wealth tax. Rich people fled instead | Fortune
fortune.com
World’s top 1% own more wealth than 95% of humanity, says Oxfam
oxfam.org
The Chair weighs what survives and returns one actionable verdict.
Chair verdict
No formal enterprise-harness standard exists yet. Harness quality, not model choice, is what determines the outcome.
Panel agreement
3 of 3 positions in
Chair confidence
Held at medium because three claims rest on vendor material alone.
The disagreement is about confidence, not the conclusion - Sceptic wants the adoption claims rated Medium.
Choose the depth
By default Dr Moot chooses the right process: Solo answers directly, Review improves one answer, Debate tests opposing sides, and Moot convenes the full panel.
One model answers directly for quick, low-stakes questions.
One answer is drafted, independently challenged, revised and judged.
Affirmative and Negative argue opposing sides and a Chair decides.
Generator, Sceptic and Specialist deliberate toward one chaired verdict.
Built for consequential work
Consensus check round 1
CompletePanel did not reach a stance
91%CONTESTED · 1
Generator's analogy that ‘taxation already limits property rights’ fully rebuts the libertarian objection that a near-100% marginal cap is categorically different from ordinary taxation.
63%*AGREED BUT UNRESOLVED · 3
Whether Generator's ‘ethically defensible in principle’ claim should be explicitly labeled as Rawlsian-leaning rather than presented as framework-neutral (Sceptic raises this as a load-bearing unsupported assumption).
AGREED AND RESOLVED · 4
A hard maximum wealth cap is ethically permissible in principle under at least some normative frameworks.
Market entry, build versus buy, pricing and operating choices.
Find one-sided clauses, unusual terms and costly omissions.
Challenge architecture and code instead of politely summarising it.
Stress-test claims, sources and counter-arguments before publishing.
Hast du irgendwelche Fragen? Wir sind für dich da.
Convene the panel
One prompt in. The strongest surviving answer out.
Start a mootSei ganz vorne dabei, wenn Dr Moot die Türen öffnet - wir senden dir eine Einladung, sobald Plätze frei werden.