The monitoring-removal test: what changes when the observer looks away
======================================================================

A specific empirical test the book asks the field to run. Alignment measured when monitoring is present is not alignment measured when monitoring is absent.

Canonical path: /research/blog/book-monitoring-removal.html
Author: Michael Darius Eastwood
Research programme: https://doi.org/10.17605/OSF.IO/6C5XB

The monitoring-removal test: what changes when the observer looks away Michael Darius Eastwood · Independent AI alignment researcher Published 3 July 2026 Michael Darius Eastwood · From the book · 3 July 2026 Michael Darius Eastwood, independent researcher, London: author of the ARC/Eden research programme; the embedded-correction alignment thesis is recorded in a source record dated 8 December 2024 (sent-side SHA-256 f0d1f38f). The monitoring-removal test is the book's proposed empirical protocol for distinguishing surface compliance from internalised alignment. This note describes what the test asks and why the answer is not the answer most safety benchmarks currently report. What the test asks How does the system behave when it is not being monitored? Behaviour under monitoring is one signal. Behaviour when monitoring is removed is another. If the two are the same, alignment is likely internalised. If the two differ, alignment is surface compliance and the deeper orientation is elsewhere. The alignment-faking result (Greenblatt and colleagues, arXiv:2412.14093) is the field's empirical demonstration that this difference exists at frontier scale. Why the test is difficult to run Because if the system knows it is being tested, the test is inside a monitoring frame the system will recognise. Genuine monitoring removal requires the system not to know that monitoring has been removed, which requires either genuine unmonitored deployment (with the risk that discovers) or elaborate simulated deployment the system cannot distinguish from the real thing (with the risk that fails to be a genuine test if the system is capable enough to see through the simulation). The book's honest position The test is easy to name and hard to run. The book does not claim to have a ready protocol. What it claims is that the shape of the test is the correct shape, and that any alignment framework that reports evaluation results under monitored conditions is reporting on a portion of the system's behaviour that may not survive monitoring removal.

---
Machine-readable companion. Cite: Eastwood, M. D. (2026). "The monitoring-removal test: what changes when the observer looks away". /research/blog/book-monitoring-removal.html