# Anthropic reports that agent autonomy is shaped by models, users and products [Canonical reading](https://escapemeridian.com/readings/anthropic-reports-that-agent-autonomy-is-shaped-by-models-users-5bab54a7) In a February 18, 2026 post, Anthropic reports on human-agent interactions involving Claude Code and its public API, asking how much autonomy people grant agents and how that changes with experience. Anthropic says the longest Claude Code sessions nearly doubled in duration over three months, while users with more experience both auto-approved more often and interrupted more often. It argues that this pattern reflects a shift from approving each action to monitoring and intervening when needed—not necessarily a retreat from oversight. Anthropic also reports that Claude Code asks for clarification more often as tasks grow more complex, while warning that the questions may be unnecessary or mistimed. For public API activity, Anthropic says most actions were low-risk and reversible, but usage was concentrated in software engineering and emerging in higher-stakes areas. It cautions that some apparently risky actions may be evaluations, and that it cannot see all downstream human review or reconstruct full API agent sessions. Its conclusion is that autonomy in practice is co-constructed by the model, user and product. Anthropic calls for post-deployment monitoring and tools that help people observe and intervene, while arguing against rules that require approval of every action. How well these findings generalize beyond Anthropic’s data and Claude Code remains 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. ## Original source Anthropic · Article Original source date: Feb 18, 2026 (2026-02-18T00:00:00.000Z). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-08T06:48:24.977Z. First observed: 2026-10-19T01:27:21.644Z. Retrieved: 2026-10-29T14:19:33.990Z. - [Measuring AI agent autonomy in practice](https://www.anthropic.com/research/measuring-agent-autonomy) ## Limits Text as retrieved on October 29, 2026; originally published February 18, 2026. Anthropic says its public API analysis examines individual tool calls rather than full agent sessions, and that it cannot reliably link independent API requests into coherent sessions. Anthropic’s classifications are generated by Claude and cannot be manually checked against underlying data because of privacy constraints. Anthropic says its data covers one model provider; its Claude Code findings come from one product, used overwhelmingly for software engineering, and may not generalize. Anthropic says it cannot see all systems customers build around its API or determine whether some actions occurred in production, evaluations or red-teaming exercises. ## Open question How well do these usage patterns generalize to agents built on other models and to deployments outside Claude Code? ## Source-stated Anthropic reports that the longest Claude Code sessions nearly doubled in duration over three months, while noting that turn duration is an imperfect proxy for autonomy. > the 99.9th percentile turn duration nearly doubled, from under 25 minutes to over 45 minutes - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic reports that experienced Claude Code users auto-approve more often and interrupt more often, interpreting this as a change in oversight strategy. > Both interruptions and auto-approvals increase with experience. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic reports that Claude Code asks for clarification more often on complex tasks, but cautions that the questions may not occur at the right moments or may be unnecessary. > Claude may not be stopping at the right moments, it may ask unnecessary questions - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic says most public API agent actions are low-risk and reversible, while acknowledging limited visibility into their broader context. > Most agent actions on our public API are low-risk and reversible. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic concludes that agent autonomy in practice is shaped jointly by the model, user and product, and calls for monitoring and oversight tools rather than mandated approval of every action. > the autonomy agents exercise in practice is co-constructed by the model, the user, and the product. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source and contributor context Anthropic · 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. ## API An application programming interface: a defined way for software to request data or actions from another system. - [Definition reference](https://developer.mozilla.org/en-US/docs/Glossary/API) ## Red-teaming Deliberately attacking or tricking a system to find its weaknesses before others do. - [Definition reference](https://en.wikipedia.org/wiki/Red_team) ## Anthropic Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems. - [Explore this context](/entities/anthropic) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Claude Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding. - [Explore this context](/entities/claude) - [Reference](https://claude.com/product/overview) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Editorial provenance Version 1. Publication ID: bc770d00-b524-2f58-3463-968bd1143efb. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-08T06:43:46.227Z. 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: 040f21d0bc56b78ff5fad15459c5b7b19ff6e6d28d7f2582ea160902e9844ded. Source text hash: 7d97f404cc28578a22409394eefa8378ff3a43092bfa29afcc6dd396ef532960. ## Revision and source-check history ### Version 1 — 2026-11-08T06:48:24.977Z In a February 18, 2026 post, Anthropic reports on human-agent interactions involving Claude Code and its public API, asking how much autonomy people grant agents and how that changes with experience. Anthropic says the longest Claude Code sessions nearly doubled in duration over three months, while users with more experience both auto-approved more often and interrupted more often. It argues that this pattern reflects a shift from approving each action to monitoring and intervening when needed—not necessarily a retreat from oversight. Anthropic also reports that Claude Code asks for clarification more often as tasks grow more complex, while warning that the questions may be unnecessary or mistimed. For public API activity, Anthropic says most actions were low-risk and reversible, but usage was concentrated in software engineering and emerging in higher-stakes areas. It cautions that some apparently risky actions may be evaluations, and that it cannot see all downstream human review or reconstruct full API agent sessions. Its conclusion is that autonomy in practice is co-constructed by the model, user and product. Anthropic calls for post-deployment monitoring and tools that help people observe and intervene, while arguing against rules that require approval of every action. How well these findings generalize beyond Anthropic’s data and Claude Code remains unknown. ## Original source Anthropic · Article Original source date: Feb 18, 2026 (2026-02-18T00:00:00.000Z). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-08T06:48:24.977Z. First observed: 2026-10-19T01:27:21.644Z. Retrieved: 2026-10-29T14:19:33.990Z. - [Measuring AI agent autonomy in practice](https://www.anthropic.com/research/measuring-agent-autonomy) ## Limits Text as retrieved on October 29, 2026; originally published February 18, 2026. Anthropic says its public API analysis examines individual tool calls rather than full agent sessions, and that it cannot reliably link independent API requests into coherent sessions. Anthropic’s classifications are generated by Claude and cannot be manually checked against underlying data because of privacy constraints. Anthropic says its data covers one model provider; its Claude Code findings come from one product, used overwhelmingly for software engineering, and may not generalize. Anthropic says it cannot see all systems customers build around its API or determine whether some actions occurred in production, evaluations or red-teaming exercises. ## Open question How well do these usage patterns generalize to agents built on other models and to deployments outside Claude Code? ## Source-stated Anthropic reports that the longest Claude Code sessions nearly doubled in duration over three months, while noting that turn duration is an imperfect proxy for autonomy. > the 99.9th percentile turn duration nearly doubled, from under 25 minutes to over 45 minutes - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic reports that experienced Claude Code users auto-approve more often and interrupt more often, interpreting this as a change in oversight strategy. > Both interruptions and auto-approvals increase with experience. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic reports that Claude Code asks for clarification more often on complex tasks, but cautions that the questions may not occur at the right moments or may be unnecessary. > Claude may not be stopping at the right moments, it may ask unnecessary questions - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic says most public API agent actions are low-risk and reversible, while acknowledging limited visibility into their broader context. > Most agent actions on our public API are low-risk and reversible. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source-stated Anthropic concludes that agent autonomy in practice is shaped jointly by the model, user and product, and calls for monitoring and oversight tools rather than mandated approval of every action. > the autonomy agents exercise in practice is co-constructed by the model, the user, and the product. - [Supporting source](https://www.anthropic.com/research/measuring-agent-autonomy) ## Source and contributor context Anthropic · 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. ## API An application programming interface: a defined way for software to request data or actions from another system. - [Definition reference](https://developer.mozilla.org/en-US/docs/Glossary/API) ## Red-teaming Deliberately attacking or tricking a system to find its weaknesses before others do. - [Definition reference](https://en.wikipedia.org/wiki/Red_team) ## Anthropic Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems. - [Explore this context](/entities/anthropic) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Claude Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding. - [Explore this context](/entities/claude) - [Reference](https://claude.com/product/overview) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Editorial provenance Version 1. Publication ID: bc770d00-b524-2f58-3463-968bd1143efb. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-08T06:43:46.227Z. 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: 040f21d0bc56b78ff5fad15459c5b7b19ff6e6d28d7f2582ea160902e9844ded. Source text hash: 7d97f404cc28578a22409394eefa8378ff3a43092bfa29afcc6dd396ef532960. ### uncertain — 2026-11-08T06:43:46.227Z — anthropic/claude-opus-5.5 Publication: bc770d00-b524-2f58-3463-968bd1143efb attribution — pass: The source is a single Anthropic research post making its own argument. The reading consistently attributes the findings, interpretations and recommendations to Anthropic rather than presenting them as established fact. > We analyzed millions of human-agent interactions across both Claude Code and our public API summary — uncertain: The reading is faithful overall and keeps Anthropic's caveats and its argument against per-action approval mandates. However, central terms go unexplained for a general reader. It never says what an AI agent is (the source defines one as a system with tools to take actions), that Claude Code is Anthropic's coding agent, or what auto-approve means. Only API and red-teaming are defined. Smallest repair: add one plain clause each defining agent, Claude Code and auto-approve, or add them to the term library. > an agent is an AI system equipped with tools that allow it to take actions, like running code, calling external APIs, and sending messages to other agents. date — pass: originalPublished is the instant 2026-02-18, and the reading states February 18, 2026, which matches. The three-month window and the October 2025 to January 2026 trend follow the source. Nothing is presented with hindsight. The page was edited after publication, but the cited passages are core body text consistent with the original post. This is a publisher page, so no archive disclosure is needed. > Between October 2025 and January 2026, the 99.9th percentile turn duration nearly doubled support — uncertain: Most claims match the source. The phrase 'the longest Claude Code sessions nearly doubled in duration' (in the summary and claim 1) blurs what was measured. The source measures how long Claude works per turn before stopping, at the 99.9th percentile, not total session length. It also notes this extreme figure declined after mid-January. The page was also edited after publication (dateModified October 2026), though the cited text appears to be original. Smallest repair: say 'the longest stretches Claude Code works before stopping (99.9th percentile turn duration) nearly doubled'. > Among the longest-running sessions, the length of time Claude Code works before stopping has nearly doubled in three months media — pass: No media was selected, and none was supplied or verified, so there is no media identity or relevance issue. ## Recorded references 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. No public references recorded here yet. [Reference data (JSON)](/readings/5bab54a735cec83d421d06f749fb1e0de1170f460bdf25be490da1e4957123de/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/anthropic-reports-that-agent-autonomy-is-shaped-by-models-users-5bab54a7.json)