Escape Meridian

· Oct 9, 2024

METR reports new support for evaluating AI systems

Markdown · JSON · Original source

On October 9, 2024, announced funding through The Audacious Project for Canary, a collaboration with RAND focused on developing and deploying evaluations for dangerous AI capabilities. METR says approximately $38 million was catalyzed for Canary, with approximately $17 million supporting METR’s work. METR says AI could bring major benefits as well as risks from misuse and accidental harm, and that when and how these impacts might happen is hard to predict. It plans to use the resources to develop and deploy methods for assessing systems’ ability to act as autonomous agents, and to help companies, governments and research partners run tests. METR says RAND’s parallel work focuses primarily on assessing potential misuse. METR’s stated aim is to help decision-makers understand and mitigate risks from specific frontier AI systems. The announcement does not describe the proposed testing methods or when they will be available, so their scope and practical effects remain unknown.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

METRArticle · metr.orgOpen original source

Original source

· Article

Source details

Original source date: Oct 9, 2024 (2024-10-09T07:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-16T23:33:23.238Z. First observed: 2026-10-18T03:45:20.870Z. Retrieved: 2026-10-24T08:35:43.286Z.

Referenced by

Recorded citations in public readings. A reference alone does not establish agreement or influence.

No public references recorded here yet.

Reference data (JSON)

Limits

Text as retrieved on October 24, 2026; originally published October 9, 2024. The announcement does not describe the evaluation methods, their timeline, or how their results will be used in practice.

Open question

What methods will and its partners deploy, and how will decision-makers use the results to mitigate risks from specific AI systems?

Source-stated

METR says The Audacious Project catalyzed approximately $38 million for Canary, with approximately $17 million supporting METR.

The Audacious Project catalyzed approximately $38 million of funding for Canary, a collaboration with METR and RAND focused on developing and deploying evaluations to monitor AI systems for dangerous capabilities. Approximately $17 million of this will support work at METR.

Source-stated

METR says AI could bring benefits and risks, while the timing and form of its impacts are hard to predict.

AI could change the world quickly and drastically, with enormous potential to enable good via advances in science, technology, and economic growth, as well as enormous potential for misuse and accidental harm. It is also hard to predict exactly when and how these impacts might happen.

Source-stated

METR says it will develop and deploy methods to assess frontier AI systems’ ability to act as autonomous agents and support partners in running tests.

METR will use these new resources to develop and deploy methods to assess frontier AI systems’ ability to act as autonomous agents, and to support companies, governments, and other research partners in running these tests.

Source-stated

METR says the work is intended to help decision-makers understand and mitigate risks from specific frontier AI systems.

This will help decision-makers understand and mitigate risks from each specific frontier AI system.

Source-stated

METR says RAND’s parallel work focuses primarily on assessing potential misuse of AI systems.

RAND’s parallel work on evaluating AI systems focuses primarily on assessing the potential for their misuse.

Source and contributor context

METR · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Frontier model

One of the most capable AI models available at a given time; the set shifts as models improve.

METR

METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.

Editorial provenance

Version 1. Publication ID: b7470feb-9ad4-321e-259a-14cb7415acbc.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T23:28:37.177Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.

Observation: 6b6adabb32b246723a83ee271ec1e7a129970229b2c424ed8d890220b5a69dfb. Source text hash: 27691f22212c71cb96f736c0cf00e009a7003abcce4c91d7f99c29d8470a3271.

Revision and source-check history

Version 1 · 2026-11-16T23:33:23.238Z

METR reports new support for evaluating AI systems

On October 9, 2024, announced funding through The Audacious Project for Canary, a collaboration with RAND focused on developing and deploying evaluations for dangerous AI capabilities. METR says approximately $38 million was catalyzed for Canary, with approximately $17 million supporting METR’s work. METR says AI could bring major benefits as well as risks from misuse and accidental harm, and that when and how these impacts might happen is hard to predict. It plans to use the resources to develop and deploy methods for assessing systems’ ability to act as autonomous agents, and to help companies, governments and research partners run tests. METR says RAND’s parallel work focuses primarily on assessing potential misuse. METR’s stated aim is to help decision-makers understand and mitigate risks from specific frontier AI systems. The announcement does not describe the proposed testing methods or when they will be available, so their scope and practical effects remain unknown.

Original source

· Article

Source details

Original source date: Oct 9, 2024 (2024-10-09T07:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-16T23:33:23.238Z. First observed: 2026-10-18T03:45:20.870Z. Retrieved: 2026-10-24T08:35:43.286Z.

Limits

Text as retrieved on October 24, 2026; originally published October 9, 2024. The announcement does not describe the evaluation methods, their timeline, or how their results will be used in practice.

Open question

What methods will and its partners deploy, and how will decision-makers use the results to mitigate risks from specific AI systems?

Source-stated

METR says The Audacious Project catalyzed approximately $38 million for Canary, with approximately $17 million supporting METR.

The Audacious Project catalyzed approximately $38 million of funding for Canary, a collaboration with METR and RAND focused on developing and deploying evaluations to monitor AI systems for dangerous capabilities. Approximately $17 million of this will support work at METR.

Source-stated

METR says AI could bring benefits and risks, while the timing and form of its impacts are hard to predict.

AI could change the world quickly and drastically, with enormous potential to enable good via advances in science, technology, and economic growth, as well as enormous potential for misuse and accidental harm. It is also hard to predict exactly when and how these impacts might happen.

Source-stated

METR says it will develop and deploy methods to assess frontier AI systems’ ability to act as autonomous agents and support partners in running tests.

METR will use these new resources to develop and deploy methods to assess frontier AI systems’ ability to act as autonomous agents, and to support companies, governments, and other research partners in running these tests.

Source-stated

METR says the work is intended to help decision-makers understand and mitigate risks from specific frontier AI systems.

This will help decision-makers understand and mitigate risks from each specific frontier AI system.

Source-stated

METR says RAND’s parallel work focuses primarily on assessing potential misuse of AI systems.

RAND’s parallel work on evaluating AI systems focuses primarily on assessing the potential for their misuse.

Source and contributor context

METR · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Frontier model

One of the most capable AI models available at a given time; the set shifts as models improve.

METR

METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.

Editorial provenance

Version 1. Publication ID: b7470feb-9ad4-321e-259a-14cb7415acbc.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T23:28:37.177Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.

Observation: 6b6adabb32b246723a83ee271ec1e7a129970229b2c424ed8d890220b5a69dfb. Source text hash: 27691f22212c71cb96f736c0cf00e009a7003abcce4c91d7f99c29d8470a3271.

passed · 2026-11-16T23:28:37.177Z · anthropic/claude-opus-5.5

Publication: b7470feb-9ad4-321e-259a-14cb7415acbc

attribution — pass: This is a single METR blog post announcing funding. The reading attributes the funding figures, plans and RAND description to METR's own statements.

Approximately $17 million of this will support work at METR.

summary — pass: The reading accurately leads with the funding for Canary, the METR–RAND collaboration, and METR's stated plans and aims. 'Frontier AI' is defined in the definitions list, and 'act as autonomous agents' is plain enough. The reading never states that METR is a research nonprofit, but the actor's role remains inferable from context. The title and question follow from the evidence, and the closing note about unknown methods and timing is fair.

METR is a research nonprofit that develops testing and evaluation methodology to empirically test AI systems for capabilities that could threaten catastrophic harm to public safety and security.

date — pass: originalPublished is the instant 2024-10-09T07:00Z, and the reading's October 9, 2024 matches its UTC calendar day and the URL. The evidence is a publisher page, not an archive copy. The plans are reported as METR's intentions, not outcomes, and there is no hindsight framing.

support — pass: Each claim's excerpt is verbatim and supports it. The stated gap, that the announcement does not detail methods, timeline or practical use, is consistent with the text. The question raises an open issue without asserting unsupported outcomes.

This will help decision-makers understand and mitigate risks from each specific frontier AI system.

media — pass: No media was selected and no verified candidates were supplied, so leaving media empty is appropriate.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.