Paul Roebuck
Behavioural researcher and curator of The Ostracon. Through SHaDS™, advisory work, writing and talks, I help people and organisations understand the behavioural consequences of AI.
14 million words written with AI since December 2022 — 1.4 million of them mine, across 2,987 conversations. Counted 30 September 2026.
Five behaviours a system makes when it can’t say “I don’t know.”
Explore → Peers & boardsPeer-facing papers on human–AI collaboration — plus governance thinking and thought-inspiration for organisations.
Read → The bookThe book behind the framework.
Read → Instrument · in betaReads behaviour under pressure — Hunter ↔ Farmer, and who’s mis-seated. In private calibration now.
Coming soon · join the beta → Level 2 instrument · in progressAim your AI at your My Documents or your repository, and see what my Excavator can surface and my other workers reveal to you — what you ask AI, how you chunk up and drill down, where it quietly defends a gap. Additional Intelligence, not agents: you stay the mind in the room. Includes development of your own governance framework.
Register interest → SHaDS™ for organisationsThe free checker warns you what to check. A governance instance goes further — your own version-controlled checks, used the same way across every team, setting what your organisation won’t let through.
See the governance standard →Twenty years in the room, now in the record — advisory, keynotes and the behavioural side of your AI rollout.
Try it
Don’t take our word for it. Copy this into your own AI — ChatGPT, Claude, Gemini — and it’ll explain SHaDS™, walk you through it, and check itself as it goes.
Pastes a full brief. Your AI reads it, then offers you four ways in.