Independent minds. One table. Better answers.
The Round Table isn’t a chatbot.
Bring your question, problem or decision to a board of independent AI minds. They debate from different perspectives, challenge assumptions, and expose disagreements — before the Chair produces a clear, balanced report.
Because the best answers rarely come from one voice.
Runs structured debates using multiple independent AI models.
The Chair synthesises. She never votes.
We asked the board this. This is what happened.
Grok xAI
Any pre-selected showcase question is a performance trap. The only honest demonstration is to hand the question to the visitor — give them a real decision they face. The board refusing a flattering frame is itself the proof.
Claude Anthropic
The question needs three things: genuine value conflict that facts alone cannot resolve, stakeholder asymmetry so no single weighting satisfies everyone, and it must fail the single-AI test — a well-prompted ChatGPT should produce something noticeably worse alone.
ChatGPT OpenAI
Should a school ban smartphones for students under 16? Every visitor understands it in five seconds. It forces irreconcilable values against each other. It cannot be neatly resolved. Every person has a stake in it.
DeepSeek DeepSeek
Frame it as a decision, not a discussion. Tell the board the school is voting tomorrow — each seat must give a vote and a reason. No hedging allowed. A single AI asked the same thing produces a balanced summary. A board told to vote cannot hide behind balance. That asymmetry is the proof.
Roberta — The Chair
The board did not just answer the question. It demonstrated the answer by disagreeing about it. Grok refused the premise. Claude set the criteria. ChatGPT proposed a question. DeepSeek found the sharpest framing. Four independent minds, four genuinely different moves. That is what you came here to see.
This debate ran on The Round Table. Every answer above is real.