Can We Govern a Race That Won’t Stop?

Rules are difficult when companies and states believe that slowing down could leave them behind.

The paired portraits and credits appear on the series home. This is Martin’s independent commentary on Kohler’s published work. Alan Kohler has not reviewed or endorsed it.

Kohler’s documented work

What he has argued

Kohler’s writing on AI governance points to a coordination problem: warnings about risk do not automatically make restraint achievable when businesses and governments face competitive pressure. Regulation must contend with incentives, uneven enforcement and the fear that another actor will move first. His argument for limiting or banning autonomous AI sits inside this wider problem of how to govern a technology that crosses organisations and borders.

Martin’s response

Where I agree, and where I go further

I see the same dilemma in post-agentic questions. It is not enough to imagine a benevolent system or a wise regulator; we have to ask who can enforce limits, how people can contest decisions and what happens when actors have different goals. A ban may be one proposal, but the public discussion also needs to spell out scope, definitions, accountability and what remains permitted.

Open question

What I still want to understand

Which controls can be verified and enforced across a competitive market? What should be prohibited outright, what could be licensed, and who gets a voice in setting those boundaries?

Research trail

Sources

Published 18 October 2026 · Commentary by Martin Davies-Roundhill · Contact End Up Here