Would even an AI disaster on the scale of Hiroshima be enough to make humankind protect itself? I fear not | Timothy Garton Ash
A columnist writing from Silicon Valley argues that AI development is outpacing humanity's ability to govern it, and that even a catastrophic AI-related disaster may not be sufficient to prompt meaningful collective action. The piece centers on the concept of recursive self-improvement, the point at which AI systems begin training successive versions of themselves.
Why this matters: The honest concern here is not science fiction. It is a governance gap: the people building these systems are moving faster than any law, treaty, or institution designed to check them. Recursive self-improvement, if it arrives, is not a software update. It is a point where human oversight becomes harder to apply, not easier. The columnist's worry is not just that something bad might happen. It is that something bad might happen and we still would not act. That is the part worth sitting with.
Who should care: AI governance · Lawyers · Administrators · General readers · Policy
This summary is AI-assisted and may contain errors. It is an original briefing to help you gauge significance quickly — not a reproduction of the source. Always read the linked original before relying on it. See our methodology.