The BaseEssayLessWrong
Control monitors need alignment assessments too
In a LessWrong post, Fabien Roger argues that AI systems used to monitor other AI agents often fail because of their own alignment problems, not weak capability, so monitors need alignment assessments much as agents do. The post lists possible research projects.
Why it matters
The same thing, for different readers.
For lawmakers
Fabien Roger's post suggests that risk reports which rely on AI monitors to catch agent misbehaviour should also argue that the monitors themselves are aligned.
For researchers
Fabien Roger argues that red-teaming monitors for capability is not enough, and proposes projects such as building a misaligned-monitor model organism and checking monitors' reasoning for signs of misalignment.
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“When misalignment monitors fail, it is already often because of alignment failures, not because they are too weak. I expect this to continue in the future.”
“In the future, the most dangerous monitor failure modes will likely also be alignment issues.”
“But getting evidence that monitors are aligned enough can still be hard, because the situations where the agent does sth bad are way more likely to be the ones where the monitor is also wrong.”
The record
- Questions
- Can AI systems act beyond what people intended or allowed? · Are AI companies putting stronger safeguards into practice?
- People
- Fabien Roger (Wrote it)
- What we read
- We read the whole source.
- Found by
- The Radar, run of October 25, 2026, via
lesswrong(tier 1: the source’s own channel or an official record) - Cite this
- LessWrong, “Control monitors need alignment assessments too”, October 24, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/lesswrong-2026-10-24-control-monitors-alignment-assessments