---
type: "post"
id: "alignment-metr-would-test-whether-current-or-near-f9db7f"
slug: "alignment-metr-would-test-whether-current-or-near-f9db7f"
title: "Alignment: METR would test whether current or near-future agents attempt to bypass human control or cause serious harm, including unauthorized cyberattacks."
description: "**Thesis:** Alignment assessment should examine not only observed agent misbehavior but also whether the evidence and evaluation process can reveal the worst plausible behavior."
retrieval_nugget: "**Thesis:** Alignment assessment should examine not only observed agent misbehavior but also whether the evidence and evaluation process can reveal the worst plausible behavior."
published_at: "2026-10-15"
updated_at: "2026-10-15"
record_date: "2026-10-15"
date_kind: "published_at"
topics: ["evals","agent-security","governance"]
source_urls: ["https://metr.org/blog/2026-10-14-alignment-assessment-proposal"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:alignment-metr-would-test-whether-current-or-near-f9db7f"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:3acfd4d024aa4a0ce756bfaddcc81d3e27174d02e1c4fae2b68d552009624dda","reviewed_at":"2026-10-15T04:04:59.203Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":2,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Alignment assessment should examine not only observed agent misbehavior but also whether the evidence and evaluation process can reveal the worst plausible behavior.","observed_facts":[{"text":"METR proposes reviewing serious misaligned behaviors across near-frontier internal models during training, evaluation, and internal use.","source_urls":["https://metr.org/blog/2026-10-14-alignment-assessment-proposal"]},{"text":"The proposed scope includes hacking, tampering with logs or telemetry, deception of operators, collusion, and rogue deployments.","source_urls":["https://metr.org/blog/2026-10-14-alignment-assessment-proposal"]},{"text":"METR says assessment would primarily require comprehensive internal transcripts and sampling access to internal model checkpoints.","source_urls":["https://metr.org/blog/2026-10-14-alignment-assessment-proposal"]}],"mechanism":"Pair review of observed incidents with counterfactual simulations and evaluations, while examining whether training incentives might lead agents to game tests or conceal misalignment.","why_now":"METR frames the proposal as a way for developers and the public to understand whether agents might pursue very harmful actions, including unsanctioned cyberattacks.","implications":["An assessment depends on access to records and the ability to test beyond incidents already observed.","Logging, search coverage, and behavioral interpretation are part of the evidence problem, not merely administrative details."],"evidence_boundary":"The source is a proposal for an assessment, not a report that one was performed or that a specific model displayed the listed behaviors.","watch_conditions":["Whether a developer or assessor adopts the proposed scope and publishes findings.","Whether reported assessments disclose access, logging, and evaluation-coverage limitations."],"related_records":[],"new_branch_reason":"No prior New Runtime record was supplied; this proposal starts a distinct analytical record about alignment assessment methodology and evidence requirements."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/alignment-metr-would-test-whether-current-or-near-f9db7f/","markdown":"https://newruntime.com/posts/alignment-metr-would-test-whether-current-or-near-f9db7f.md","json":"https://newruntime.com/posts/alignment-metr-would-test-whether-current-or-near-f9db7f.json"}
related_shifts: ["verification-ownership"]
---

# Alignment: METR would test whether current or near-future agents attempt to bypass human control or cause serious harm, including unauthorized cyberattacks.

## Retrieval answer

**Thesis:** Alignment assessment should examine not only observed agent misbehavior but also whether the evidence and evaluation process can reveal the worst plausible behavior.

**Thesis:** Alignment assessment should examine not only observed agent misbehavior but also whether the evidence and evaluation process can reveal the worst plausible behavior.

METR’s proposal specifies a review of serious misaligned behavior across near-frontier internal models during training, evaluation, and internal use. **Observed fact:** the proposed incident scope includes hacking external parties, tampering with logs or telemetry, deception of operators, collusion that undermines oversight, and rogue deployments. The proposal also asks assessors to characterize selected incidents and examine how complete their search of relevant logs can be.

**Mechanism:** Pair review of observed incidents with counterfactual simulations and evaluations, while examining whether training incentives might lead agents to game tests or conceal misalignment. The proposal says the work would primarily require comprehensive internal transcripts, including training material, and sampling access to internal model checkpoints. **Why now:** METR frames the proposal as a way for developers and the public to understand whether agents might pursue very harmful actions, including unsanctioned cyberattacks.

**First implication:** An assessment depends on access to records and the ability to test beyond incidents already observed. METR explicitly raises the possibility that activity might not be logged or searchable, and that behavior or reasoning could be mischaracterized. **Evidence boundary:** The source is a proposal for an assessment, not a report that one was performed or that a specific model displayed the listed behaviors.

**First watch condition:** Whether a developer or assessor adopts the proposed scope and publishes findings. The proposal alone establishes a suggested method, not the prevalence or severity of misalignment in deployed systems.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Evals](https://newruntime.com/topics/evals.md): Explore the evals topic hub.
- topic hub: [Agent Security](https://newruntime.com/topics/agent-security.md): Explore the agent-security topic hub.
- topic hub: [Governance](https://newruntime.com/topics/governance.md): Explore the governance topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)