Thoughts on "safety bad bc warning shots good"

In a LessWrong post, Cleo Nardo examines the Warning Shot Argument, which holds that making today's AI systems safer could raise catastrophic risk by reducing the minor harms that wake people up. The excerpt sets out the argument and begins an OpenAI example.

Why it matters

The same thing, for different readers.

What it means for you

Cleo Nardo asks whether preventing small AI-caused harms now could leave the public less alert to bigger dangers later.

For lawmakers

The argument, as Cleo Nardo states it, claims lawmakers and labs act on catastrophic risk only after non-catastrophic incidents such as cybercrime.

For researchers

Cleo Nardo frames the question as a thought experiment: whether a button making current models more aligned should be pressed if it does not change future models.

The question it raises

Cleo Nardo raises the ethical tension of tolerating smaller harms today in exchange for a possible reduction in catastrophic risk later.

The record

Questions
Can AI systems act beyond what people intended or allowed? · Are AI companies putting stronger safeguards into practice?
People
Cleo Nardo (Wrote it)
What we read
We have read only the description the publisher gives. No quotations until someone reads the source itself.
Found by
The Radar, run of November 17, 2026, via lesswrong (tier 1: the source’s own channel or an official record)
Cite this
LessWrong, “Thoughts on "safety bad bc warning shots good"”, November 17, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/lesswrong-2026-11-17-safety-bad-warning-shots-good
Our notesLight lane: written from the triage excerpt only, which is cut off; Cleo Nardo's own conclusion is not visible, so the post's stance is not stated here.