The BaseEssayAI Alignment Forum
Monitorability evaluations and policy proposals
On the AI Alignment Forum, evhub sets out four types of monitorability, argues trust in monitors should rest on auditing evaluations, and proposes a methodology for measuring them. The post also backs both binding red lines and ongoing limits on lab changes that reduce monitorability. evhub says these are personal views.
Why it matters
The same thing, for different readers.
For lawmakers
evhub proposes two kinds of rule: red lines on how unmonitorable models can be, plus checks on model changes that worsen monitorability before any red line is crossed.
For researchers
evhub splits monitorability into control, training, auditing and diffuse types and proposes measuring the first three against the best deployed trusted monitor, with real elicitation effort.
The question it raises
Monitoring a model with a model raises the question of whose judgment can be trusted when the watched system might be misaligned.
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.
“A basic fact that is true about at least all of the non-diffuse forms of monitorability is that you should be able to in some sense get a perfect score by just using the same model as the monitor and policy.”
“The big thing I'd like to see is a combination of a) red lines for absolute levels of unmonitorability that labs should not cross and b) measures that constrain the degree to which labs can do things that degrade monitorability even before reaching those red lines.”
“Labs should not be able to internally deploy models that can reliably evade detection when attempting dangerous sabotage actions like breaking out of a sandbox, poisoning data, or sabotaging an alignment experiment.”
The record
- Questions
- Can AI systems act beyond what people intended or allowed? · Are AI companies putting stronger safeguards into practice? · Who can demand evidence, delay a release or stop an unsafe deployment?
- People
- evhub (Wrote it) · Bronson Schoen (Mentioned)
- What we read
- We read the whole source.
- Found by
- The Radar, run of October 23, 2026, via
ai-alignment-forum(tier 1: the source’s own channel or an official record) - Cite this
- AI Alignment Forum, “Monitorability evaluations and policy proposals”, October 22, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/alignmentforum-2026-10-22-monitorability-evals-policy