The BaseEssayMETR

A proposal for better informing the public about loss-of-control risks from frontier AI development

METR proposes that frontier developers let it investigate three questions, on alignment, monitoring and AI-driven acceleration, and publish reports from staff interviews and internal data. Developers could redact material and set access. METR lists limits, including that access stays at company discretion.

Open the source ↗Top story of the week
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Today outsiders cannot tell whether AI companies still control their own internal AI systems; METR wants independent, published evidence on that.

For lawmakers

The proposal is voluntary and is not an audit; METR says it leaves the decision on what to do with the evidence to companies, governments and the public.

For researchers

Names three evidence gaps (alignment, monitoring, AI R&D acceleration) and admits the science of loss-of-control assessment is immature.

The question it raises

Asks who may see inside powerful AI development, and whether a company-chosen outside assessor can be independent.

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.

“We instead aim to get new independent evidence out to the public so that company staff, boards, external scientists, and policymakers are able to arrive at their own conclusions.”
METRMETR on what its proposed assessments are for, after saying they cannot give confident assurance of safety.
Found word for word in the source
“Companies would be able to redact any information. We intend to always publish what kinds of material information were redacted, but we could fail to reach agreement on how to do this. If so, we would publish that fact.”
METRListed among the limits of METR's proposal.
Found word for word in the source
“No company has released detailed public evidence about this, leaving significant uncertainty about whether progress is accelerating.”
METROn evidence about whether capabilities progress is accelerating through AI automation of AI R&D.
Found word for word in the source
“The science of studying AI loss of control is in its infancy. While we can provide valuable and actionable evidence to society, it is not currently possible for risk assessment exercises to provide a similar degree of assurance against loss-of-control risks that audits and regulation provide in, for example, finance.”
METRMETR's closing statement of the limits of its proposal.
Found word for word in the source

The record

Questions
Can AI systems act beyond what people intended or allowed? · Who can demand evidence, delay a release or stop an unsafe deployment? · Are AI companies putting stronger safeguards into practice?
People
Jakub Pachocki (Mentioned, Chief Scientist at OpenAI) · Paul Christiano (Mentioned, OpenAI Safety and Security Committee) · Dario Amodei (Mentioned, Anthropic CEO)
What we read
We read the whole source.
Found by
The Radar, run of October 15, 2026, via metr (tier 1: the source’s own channel or an official record)
Cite this
METR, “A proposal for better informing the public about loss-of-control risks from frontier AI development”, October 14, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/metr-2026-10-14-informing-public-loss-of-control-risks
Our notesMETR's own proposal, so the claims about incidents (OpenAI/Hugging Face and others) and the Frontier Risk Report are METR's characterisation and are not checked here. The post cites statements by Pachocki, Christiano and Amodei in a footnote; not quoted, as they are second-hand. The post mentions an in-depth alignment assessment METR is doing with Anthropic, which is also a company named in the post.