{"id":"70ab68009a46d60e64016bae7d10bdf4d7a61f7be7c62658cd019eb7688c450f","title":"METR argues for earlier transparency on frontier AI risks","argument":"METR published this preliminary writeup on June 27, 2025, saying it falls below its usual publication standard but may help decision-makers. METR argues that decision-makers may miss important evidence if they cannot see how AI models are developed and used. Public system cards may not reveal what happened during training, or risks from models kept internal. METR says early visibility matters because interventions may be most effective before developers make large investments in a model.\n\nThe writeup proposes questions about internal capabilities, possible misalignment, high-stakes model use and whether a small group could misuse powerful systems. METR sketches external risk reviews, with reports shared privately with oversight bodies and, where appropriate, aggregated and redacted for public release. It also suggests other transparency options, including external access to models.\n\nMETR warns that disclosure rules could discourage developers from finding safety problems, divert safety work or expose sensitive information. The questions have not undergone extensive iteration with frontier AI companies, so their costs, benefits and possible backfire effects remain to be assessed. METR’s aim is to make safety-relevant information available while minimizing costs and perverse incentives for developers.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 29, 2026; originally published June 27, 2025. METR labels this a preliminary writeup that falls below its usual publication standard. METR says its questions have not yet undergone extensive iteration with frontier AI companies; their costs, benefits and potential to backfire remain to be assessed. The source notes that it does not address biological weapons risks.","speaker":{"name":"METR","roleAtSource":"Publisher"},"source":{"title":"What should companies share about risks from frontier AI models?","publisher":"METR","format":"Article","url":"https://metr.org/blog/2025-06-27-risk-transparency/","publishedAt":"2025-06-27T07:00:00.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"METR says decision-makers may miss important evidence about frontier AI development when they lack visibility into how models are developed and used.","excerpt":"By default, we expect that the relevant decision-makers won’t be tracking important evidence because they won’t have visibility into how AI models are being developed and used.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR argues that information about training and development can reveal risks that may not be detectable from the final model’s output behavior or public-access evaluations.","excerpt":"To catch these issues, we need information about what happened during training and development of this model (and related models).","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR says visibility early in development matters because interventions may be most effective before a developer has made large investments in a model.","excerpt":"For both internal and external use, interventions may be most effective before an AI company has made huge investments into a particular model.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR warns that some transparency requirements could discourage developers from discovering safety issues or divert safety resources from other work.","excerpt":"Some kinds of transparency might incentivize developers to avoid finding out about important safety issues, so that they don’t have to disclose them.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR proposes that an external risk review team could produce reports for oversight bodies, with an intermediary potentially aggregating, redacting and anonymizing information for public sharing.","excerpt":"Every 3 months, an external3 “risk review” or “stress-testing” team performs a catastrophic risk investigation for a particular frontier AI developer.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR says its questions have not yet undergone extensive iteration with frontier AI companies, leaving their costs and possible backfire effects to be explored.","excerpt":"These questions have not yet undergone extensive iteration with frontier AI companies.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How could transparency requirements provide useful safety information without discouraging developers from finding problems or diverting safety work?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"METR argues for earlier transparency on frontier AI risks","summary":"METR published this preliminary writeup on June 27, 2025, saying it falls below its usual publication standard but may help decision-makers. METR argues that decision-makers may miss important evidence if they cannot see how AI models are developed and used. Public system cards may not reveal what happened during training, or risks from models kept internal. METR says early visibility matters because interventions may be most effective before developers make large investments in a model.\n\nThe writeup proposes questions about internal capabilities, possible misalignment, high-stakes model use and whether a small group could misuse powerful systems. METR sketches external risk reviews, with reports shared privately with oversight bodies and, where appropriate, aggregated and redacted for public release. It also suggests other transparency options, including external access to models.\n\nMETR warns that disclosure rules could discourage developers from finding safety problems, divert safety work or expose sensitive information. The questions have not undergone extensive iteration with frontier AI companies, so their costs, benefits and possible backfire effects remain to be assessed. METR’s aim is to make safety-relevant information available while minimizing costs and perverse incentives for developers.","question":"How could transparency requirements provide useful safety information without discouraging developers from finding problems or diverting safety work?","claims":[{"text":"METR says decision-makers may miss important evidence about frontier AI development when they lack visibility into how models are developed and used.","excerpt":"By default, we expect that the relevant decision-makers won’t be tracking important evidence because they won’t have visibility into how AI models are being developed and used.","support":"source-stated"},{"text":"METR argues that information about training and development can reveal risks that may not be detectable from the final model’s output behavior or public-access evaluations.","excerpt":"To catch these issues, we need information about what happened during training and development of this model (and related models).","support":"source-stated"},{"text":"METR says visibility early in development matters because interventions may be most effective before a developer has made large investments in a model.","excerpt":"For both internal and external use, interventions may be most effective before an AI company has made huge investments into a particular model.","support":"source-stated"},{"text":"METR warns that some transparency requirements could discourage developers from discovering safety issues or divert safety resources from other work.","excerpt":"Some kinds of transparency might incentivize developers to avoid finding out about important safety issues, so that they don’t have to disclose them.","support":"source-stated"},{"text":"METR proposes that an external risk review team could produce reports for oversight bodies, with an intermediary potentially aggregating, redacting and anonymizing information for public sharing.","excerpt":"Every 3 months, an external3 “risk review” or “stress-testing” team performs a catastrophic risk investigation for a particular frontier AI developer.","support":"source-stated"},{"text":"METR says its questions have not yet undergone extensive iteration with frontier AI companies, leaving their costs and possible backfire effects to be explored.","excerpt":"These questions have not yet undergone extensive iteration with frontier AI companies.","support":"source-stated"}],"limitations":["Text as retrieved on October 29, 2026; originally published June 27, 2025.","METR labels this a preliminary writeup that falls below its usual publication standard.","METR says its questions have not yet undergone extensive iteration with frontier AI companies; their costs, benefits and potential to backfire remain to be assessed. The source notes that it does not address biological weapons risks."],"mediaId":null,"correctionReason":null,"id":"70ab68009a46d60e64016bae7d10bdf4d7a61f7be7c62658cd019eb7688c450f","publicationId":"f38cac91-4a4e-4bbd-0a8e-9c77ca539d5a","version":1,"sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","sourceTitle":"What should companies share about risks from frontier AI models?","publisher":"METR","publishedAt":"2025-06-27T07:00:00.000Z","originalPublished":{"value":"2025-06-27T07:00:00.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-10-18T03:29:19.688Z","createdAt":"2026-11-13T09:28:23.196Z","historical":true,"entityIds":["metr"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"2dc12f665380e4889587c52120c939d6941227a0df508786854e64ea6f5ae512","rubricVersion":"unrecorded","provider":"anthropic","model":"claude-opus-5-5","reasonCodes":["bridge_admitted_by_model","second_opinion_categorical"],"confidence":1,"decidedAt":"2026-10-29T23:56:11.028Z"},"provenance":{"textHash":"a6b5b800414e10ad50eb3de788c77484332d943696f697b4363f64fdb6fb752a","observationId":"317a926640053bec97e090cc35cd5c92dafc7bcffbad4646a0eaf8b68d915516","retrievedAt":"2026-10-29T23:57:45.603Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"1d09dccf-d563-8b99-5909-da5bdc036eaa"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"7ed7183a-19a1-36e3-55d4-a9271a2e9264"}},"audit":{"status":"passed","reviewedAt":"2026-11-13T09:23:39.606Z","model":"anthropic/claude-opus-5.5"}},"canonicalUrl":"https://escapemeridian.com/readings/metr-argues-for-earlier-transparency-on-frontier-ai-risks-70ab6800","formats":{"html":"/readings/metr-argues-for-earlier-transparency-on-frontier-ai-risks-70ab6800","markdown":"/readings/metr-argues-for-earlier-transparency-on-frontier-ai-risks-70ab6800.md","json":"/readings/metr-argues-for-earlier-transparency-on-frontier-ai-risks-70ab6800.json"},"attribution":"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.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-frontier-model","name":"Frontier model","explanation":"One of the most capable AI models available at a given time; the set shifts as models improve.","referenceUrl":"https://www.gov.uk/government/publications/frontier-ai-capabilities-and-risks-discussion-paper"},{"id":"term-ai-alignment","name":"AI alignment","explanation":"Research and practice aimed at making AI systems pursue the goals their builders and users intend.","referenceUrl":"https://en.wikipedia.org/wiki/AI_alignment"}],"entities":[{"id":"metr","name":"METR","kind":"organization","context":"METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.","referenceUrls":["https://metr.org/"],"path":"/entities/metr","json":"/entities/metr.json","markdown":"/entities/metr.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"70ab68009a46d60e64016bae7d10bdf4d7a61f7be7c62658cd019eb7688c450f","revision":"f38cac91-4a4e-4bbd-0a8e-9c77ca539d5a","revisionKind":"publication-id","title":"METR argues for earlier transparency on frontier AI risks","url":"https://escapemeridian.com/readings/metr-argues-for-earlier-transparency-on-frontier-ai-risks-70ab6800"},"source":{"type":"source","id":"https://metr.org/blog/2025-06-27-risk-transparency/","revision":"a6b5b800414e10ad50eb3de788c77484332d943696f697b4363f64fdb6fb752a","revisionKind":"sha256-source-text","title":"What should companies share about risks from frontier AI models?","url":"https://metr.org/blog/2025-06-27-risk-transparency/"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"70ab68009a46d60e64016bae7d10bdf4d7a61f7be7c62658cd019eb7688c450f","title":"METR argues for earlier transparency on frontier AI risks","argument":"METR published this preliminary writeup on June 27, 2025, saying it falls below its usual publication standard but may help decision-makers. METR argues that decision-makers may miss important evidence if they cannot see how AI models are developed and used. Public system cards may not reveal what happened during training, or risks from models kept internal. METR says early visibility matters because interventions may be most effective before developers make large investments in a model.\n\nThe writeup proposes questions about internal capabilities, possible misalignment, high-stakes model use and whether a small group could misuse powerful systems. METR sketches external risk reviews, with reports shared privately with oversight bodies and, where appropriate, aggregated and redacted for public release. It also suggests other transparency options, including external access to models.\n\nMETR warns that disclosure rules could discourage developers from finding safety problems, divert safety work or expose sensitive information. The questions have not undergone extensive iteration with frontier AI companies, so their costs, benefits and possible backfire effects remain to be assessed. METR’s aim is to make safety-relevant information available while minimizing costs and perverse incentives for developers.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 29, 2026; originally published June 27, 2025. METR labels this a preliminary writeup that falls below its usual publication standard. METR says its questions have not yet undergone extensive iteration with frontier AI companies; their costs, benefits and potential to backfire remain to be assessed. The source notes that it does not address biological weapons risks.","speaker":{"name":"METR","roleAtSource":"Publisher"},"source":{"title":"What should companies share about risks from frontier AI models?","publisher":"METR","format":"Article","url":"https://metr.org/blog/2025-06-27-risk-transparency/","publishedAt":"2025-06-27T07:00:00.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"METR says decision-makers may miss important evidence about frontier AI development when they lack visibility into how models are developed and used.","excerpt":"By default, we expect that the relevant decision-makers won’t be tracking important evidence because they won’t have visibility into how AI models are being developed and used.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR argues that information about training and development can reveal risks that may not be detectable from the final model’s output behavior or public-access evaluations.","excerpt":"To catch these issues, we need information about what happened during training and development of this model (and related models).","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR says visibility early in development matters because interventions may be most effective before a developer has made large investments in a model.","excerpt":"For both internal and external use, interventions may be most effective before an AI company has made huge investments into a particular model.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR warns that some transparency requirements could discourage developers from discovering safety issues or divert safety resources from other work.","excerpt":"Some kinds of transparency might incentivize developers to avoid finding out about important safety issues, so that they don’t have to disclose them.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR proposes that an external risk review team could produce reports for oversight bodies, with an intermediary potentially aggregating, redacting and anonymizing information for public sharing.","excerpt":"Every 3 months, an external3 “risk review” or “stress-testing” team performs a catastrophic risk investigation for a particular frontier AI developer.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"},{"text":"METR says its questions have not yet undergone extensive iteration with frontier AI companies, leaving their costs and possible backfire effects to be explored.","excerpt":"These questions have not yet undergone extensive iteration with frontier AI companies.","url":"https://metr.org/blog/2025-06-27-risk-transparency/","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How could transparency requirements provide useful safety information without discouraging developers from finding problems or diverting safety work?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"METR argues for earlier transparency on frontier AI risks","summary":"METR published this preliminary writeup on June 27, 2025, saying it falls below its usual publication standard but may help decision-makers. METR argues that decision-makers may miss important evidence if they cannot see how AI models are developed and used. Public system cards may not reveal what happened during training, or risks from models kept internal. METR says early visibility matters because interventions may be most effective before developers make large investments in a model.\n\nThe writeup proposes questions about internal capabilities, possible misalignment, high-stakes model use and whether a small group could misuse powerful systems. METR sketches external risk reviews, with reports shared privately with oversight bodies and, where appropriate, aggregated and redacted for public release. It also suggests other transparency options, including external access to models.\n\nMETR warns that disclosure rules could discourage developers from finding safety problems, divert safety work or expose sensitive information. The questions have not undergone extensive iteration with frontier AI companies, so their costs, benefits and possible backfire effects remain to be assessed. METR’s aim is to make safety-relevant information available while minimizing costs and perverse incentives for developers.","question":"How could transparency requirements provide useful safety information without discouraging developers from finding problems or diverting safety work?","claims":[{"text":"METR says decision-makers may miss important evidence about frontier AI development when they lack visibility into how models are developed and used.","excerpt":"By default, we expect that the relevant decision-makers won’t be tracking important evidence because they won’t have visibility into how AI models are being developed and used.","support":"source-stated"},{"text":"METR argues that information about training and development can reveal risks that may not be detectable from the final model’s output behavior or public-access evaluations.","excerpt":"To catch these issues, we need information about what happened during training and development of this model (and related models).","support":"source-stated"},{"text":"METR says visibility early in development matters because interventions may be most effective before a developer has made large investments in a model.","excerpt":"For both internal and external use, interventions may be most effective before an AI company has made huge investments into a particular model.","support":"source-stated"},{"text":"METR warns that some transparency requirements could discourage developers from discovering safety issues or divert safety resources from other work.","excerpt":"Some kinds of transparency might incentivize developers to avoid finding out about important safety issues, so that they don’t have to disclose them.","support":"source-stated"},{"text":"METR proposes that an external risk review team could produce reports for oversight bodies, with an intermediary potentially aggregating, redacting and anonymizing information for public sharing.","excerpt":"Every 3 months, an external3 “risk review” or “stress-testing” team performs a catastrophic risk investigation for a particular frontier AI developer.","support":"source-stated"},{"text":"METR says its questions have not yet undergone extensive iteration with frontier AI companies, leaving their costs and possible backfire effects to be explored.","excerpt":"These questions have not yet undergone extensive iteration with frontier AI companies.","support":"source-stated"}],"limitations":["Text as retrieved on October 29, 2026; originally published June 27, 2025.","METR labels this a preliminary writeup that falls below its usual publication standard.","METR says its questions have not yet undergone extensive iteration with frontier AI companies; their costs, benefits and potential to backfire remain to be assessed. The source notes that it does not address biological weapons risks."],"mediaId":null,"correctionReason":null,"id":"70ab68009a46d60e64016bae7d10bdf4d7a61f7be7c62658cd019eb7688c450f","publicationId":"f38cac91-4a4e-4bbd-0a8e-9c77ca539d5a","version":1,"sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","sourceTitle":"What should companies share about risks from frontier AI models?","publisher":"METR","publishedAt":"2025-06-27T07:00:00.000Z","originalPublished":{"value":"2025-06-27T07:00:00.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-10-18T03:29:19.688Z","createdAt":"2026-11-13T09:28:23.196Z","historical":true,"entityIds":["metr"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2025-06-27-risk-transparency/","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"2dc12f665380e4889587c52120c939d6941227a0df508786854e64ea6f5ae512","rubricVersion":"unrecorded","provider":"anthropic","model":"claude-opus-5-5","reasonCodes":["bridge_admitted_by_model","second_opinion_categorical"],"confidence":1,"decidedAt":"2026-10-29T23:56:11.028Z"},"provenance":{"textHash":"a6b5b800414e10ad50eb3de788c77484332d943696f697b4363f64fdb6fb752a","observationId":"317a926640053bec97e090cc35cd5c92dafc7bcffbad4646a0eaf8b68d915516","retrievedAt":"2026-10-29T23:57:45.603Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"1d09dccf-d563-8b99-5909-da5bdc036eaa"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"7ed7183a-19a1-36e3-55d4-a9271a2e9264"}},"audit":{"status":"passed","reviewedAt":"2026-11-13T09:23:39.606Z","model":"anthropic/claude-opus-5.5"}}}],"audits":[{"publicationId":"f38cac91-4a4e-4bbd-0a8e-9c77ca539d5a","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading consistently attributes the arguments, proposals and caveats to METR, the publisher of this single blog writeup. It presents these assertions as METR's positions rather than as established facts.","excerpt":"Currently, one of our central priorities is improving the transparency of frontier AI development."},{"dimension":"summary","status":"pass","explanation":"The reading leads with METR's argument and its rationale: visibility into training, internal models and early development. It covers the question areas, the example scheme, design tradeoffs and the preliminary caveat. Frontier and misalignment are defined beside the reading. The meaning of \"system cards\" is clear enough from context as public model documentation; a tooltip would help but is not required. The title and question follow the source's own tradeoff concern.","excerpt":"Transparency beyond what is published in “system cards” when the model is publicly deployed is particularly valuable"},{"dimension":"date","status":"pass","explanation":"The reading's June 27, 2025 date matches the UTC calendar day of originalPublished. The source's \"As of June 2025\" is consistent with that date. The evidence is a publisher page, so no archive disclosure is needed. The reading does not use hindsight framing or later events, and it notes the retrieval date only as a limitation.","excerpt":"As of June 2025, the specific questions we think are most valuable for frontier AI developers to answer are:"},{"dimension":"support","status":"pass","explanation":"Each claim matches the cited text. The mention of external access to models is supported by the source's proposal to require deployment of models to external researchers. The reading implies no benefits beyond what METR states, and it keeps the backfire and competitiveness concerns intact.","excerpt":"One example would be requiring deployment of models to some minimum set of external researchers (or to the public) within some timespan after the model is developed"},{"dimension":"media","status":"pass","explanation":"No media was selected and no verified media candidates were supplied, so there is no media identity or relevance issue to assess.","excerpt":null}],"createdAt":"2026-11-13T09:23:39.606Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}