{"id":"d9f71c58c11fbed0261bd97752ffefeeba10ede69ed0e5f6cab99bb7f632048e","title":"Anthropic and Andon Labs report on AI drone capabilities and oversight","argument":"On July 24, 2026, Anthropic and Andon Labs report on tests of AI models controlling a drone to find and follow a person in an office. Their benchmark, Drone-Bench, breaks the task into subtasks, including reconstructing the space, locating and navigating the drone, detecting a person, and following them. The best-performing model exceeded the human-AI team baseline on every task except reconstruction; in a real-drone demonstration, it followed the person better than the reference algorithm but could not navigate between rooms. The authors argue that this gap matters because increasingly capable systems could make human oversight seem like a cost rather than a safeguard, despite risks to physical security and privacy. They say decisions about oversight should be deliberate, and that investment in AI alignment, governance, and safety should increase with capability. The authors also stress that these results do not establish operational capability: tests used slow-moving drones in one office floorplan and did not cover outdoor settings or large crowds. Whether models can reliably pass the baseline while operating essentially autonomously remains an open question.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 29, 2026; originally published July 24, 2026. The evaluation used slow-moving drones in one office floorplan with a limited number of people; Andon Labs did not test outdoors in large crowds. Anthropic says it was not given access to Drone-Bench; Andon Labs ran the evaluations reported here.","speaker":{"name":"Anthropic","roleAtSource":"Publisher"},"source":{"title":"Project Pilot: Can AI models fly drones?","publisher":"Anthropic","format":"Article","url":"https://www.anthropic.com/research/project-pilot","publishedAt":"2026-07-24T00:00:00.000Z","dateBasis":"Date stated on the publisher’s page."},"evidence":[{"text":"Anthropic and Andon Labs report evaluating models on a drone-based locate-and-follow task and introduce Drone-Bench.","excerpt":"we developed a new series of demonstrations and evaluations that assess AI models’ ability to use a flying drone to autonomously perform a simple locate-and-follow task","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"Andon Labs created Drone-Bench, with Anthropic consulted, and ran the reported evaluations.","excerpt":"Drone-Bench is a benchmark created by Andon Labs (in consultation with Anthropic) to test if AI agents are capable of controlling a drone for surveillance tasks.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The best-performing model exceeded the baseline on every task except reconstruction, but could not autonomously navigate between rooms in the real-drone demonstration.","excerpt":"The best performing model was Claude Fable 5, which brings the frontier past the baseline on all tasks except reconstruction.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors argue that pressure to reduce oversight may rise when models pass capability and reliability thresholds, and that oversight decisions should be deliberate.","excerpt":"there will be real pressure to treat human oversight as a cost rather than a safeguard. That is exactly why these decisions must be made deliberately","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors say the evaluation does not establish operational capability and identify limits in its test setting.","excerpt":"the drones are moving at slow speeds, we only tested in one office floorplan with a limited number of people, and Andon did not test outdoors in large crowds","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors argue that investment in AI alignment, governance, and safety should increase with the scale of capabilities.","excerpt":"the requirements for investing in AI alignment, governance, and safety increase with the scale of capabilities.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.anthropic.com/research/project-pilot","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Can models operating essentially autonomously reliably pass the human-AI team baseline?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Anthropic and Andon Labs report on AI drone capabilities and oversight","summary":"On July 24, 2026, Anthropic and Andon Labs report on tests of AI models controlling a drone to find and follow a person in an office. Their benchmark, Drone-Bench, breaks the task into subtasks, including reconstructing the space, locating and navigating the drone, detecting a person, and following them. The best-performing model exceeded the human-AI team baseline on every task except reconstruction; in a real-drone demonstration, it followed the person better than the reference algorithm but could not navigate between rooms. The authors argue that this gap matters because increasingly capable systems could make human oversight seem like a cost rather than a safeguard, despite risks to physical security and privacy. They say decisions about oversight should be deliberate, and that investment in AI alignment, governance, and safety should increase with capability. The authors also stress that these results do not establish operational capability: tests used slow-moving drones in one office floorplan and did not cover outdoor settings or large crowds. Whether models can reliably pass the baseline while operating essentially autonomously remains an open question.","question":"Can models operating essentially autonomously reliably pass the human-AI team baseline?","claims":[{"text":"Anthropic and Andon Labs report evaluating models on a drone-based locate-and-follow task and introduce Drone-Bench.","excerpt":"we developed a new series of demonstrations and evaluations that assess AI models’ ability to use a flying drone to autonomously perform a simple locate-and-follow task","support":"source-stated"},{"text":"Andon Labs created Drone-Bench, with Anthropic consulted, and ran the reported evaluations.","excerpt":"Drone-Bench is a benchmark created by Andon Labs (in consultation with Anthropic) to test if AI agents are capable of controlling a drone for surveillance tasks.","support":"source-stated"},{"text":"The best-performing model exceeded the baseline on every task except reconstruction, but could not autonomously navigate between rooms in the real-drone demonstration.","excerpt":"The best performing model was Claude Fable 5, which brings the frontier past the baseline on all tasks except reconstruction.","support":"source-stated"},{"text":"The authors argue that pressure to reduce oversight may rise when models pass capability and reliability thresholds, and that oversight decisions should be deliberate.","excerpt":"there will be real pressure to treat human oversight as a cost rather than a safeguard. That is exactly why these decisions must be made deliberately","support":"source-stated"},{"text":"The authors say the evaluation does not establish operational capability and identify limits in its test setting.","excerpt":"the drones are moving at slow speeds, we only tested in one office floorplan with a limited number of people, and Andon did not test outdoors in large crowds","support":"source-stated"},{"text":"The authors argue that investment in AI alignment, governance, and safety should increase with the scale of capabilities.","excerpt":"the requirements for investing in AI alignment, governance, and safety increase with the scale of capabilities.","support":"source-stated"}],"limitations":["Text as retrieved on October 29, 2026; originally published July 24, 2026.","The evaluation used slow-moving drones in one office floorplan with a limited number of people; Andon Labs did not test outdoors in large crowds.","Anthropic says it was not given access to Drone-Bench; Andon Labs ran the evaluations reported here."],"mediaId":null,"correctionReason":null,"id":"d9f71c58c11fbed0261bd97752ffefeeba10ede69ed0e5f6cab99bb7f632048e","publicationId":"5dcd0673-488b-826a-0be5-40e8720d9e01","version":1,"sourceUrl":"https://www.anthropic.com/research/project-pilot","sourceTitle":"Project Pilot: Can AI models fly drones?","publisher":"Anthropic","publishedAt":"2026-07-24T00:00:00.000Z","originalPublished":{"value":"2026-07-24T00:00:00.000Z","precision":"instant","basis":"backfill:article_published"},"firstObservedAt":"2026-10-20T07:11:24.016Z","createdAt":"2026-11-06T07:38:24.468Z","historical":true,"entityIds":["anthropic"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.anthropic.com/research/project-pilot","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"b05bd850b63b952339576f2db1e0d72671c97b028a6e5dd67c7760114611a506","rubricVersion":"unrecorded","provider":"typesafe","model":"jev-1.13.0","reasonCodes":["bridge_admitted_by_model"],"confidence":0.62,"decidedAt":"2026-10-29T11:17:48.636Z"},"provenance":{"textHash":"3ab9772f682396ffe10025eac156ce610760f5c4df2ec119419668cf5a6f6bbd","observationId":"f3f57687d6da867f8afe9350df183e2aa1905b603f68a7a6a3fa21ed91482ad1","retrievedAt":"2026-10-29T11:19:31.016Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"3f15a897-fccb-a12d-a952-d4aa6c3646bc"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"1a4ee96c-bb26-94cd-837b-3e3ea93f0176"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-06T07:34:26.882Z","model":"anthropic/claude-opus-5.5"}},"canonicalUrl":"https://escapemeridian.com/readings/anthropic-and-andon-labs-report-on-ai-drone-capabilities-and-d9f71c58","formats":{"html":"/readings/anthropic-and-andon-labs-report-on-ai-drone-capabilities-and-d9f71c58","markdown":"/readings/anthropic-and-andon-labs-report-on-ai-drone-capabilities-and-d9f71c58.md","json":"/readings/anthropic-and-andon-labs-report-on-ai-drone-capabilities-and-d9f71c58.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-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"},{"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":"anthropic","name":"Anthropic","kind":"organization","context":"Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.","referenceUrls":["https://www.anthropic.com/news","https://www.anthropic.com/research/team/economics"],"path":"/entities/anthropic","json":"/entities/anthropic.json","markdown":"/entities/anthropic.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"d9f71c58c11fbed0261bd97752ffefeeba10ede69ed0e5f6cab99bb7f632048e","revision":"5dcd0673-488b-826a-0be5-40e8720d9e01","revisionKind":"publication-id","title":"Anthropic and Andon Labs report on AI drone capabilities and oversight","url":"https://escapemeridian.com/readings/anthropic-and-andon-labs-report-on-ai-drone-capabilities-and-d9f71c58"},"source":{"type":"source","id":"https://www.anthropic.com/research/project-pilot","revision":"3ab9772f682396ffe10025eac156ce610760f5c4df2ec119419668cf5a6f6bbd","revisionKind":"sha256-source-text","title":"Project Pilot: Can AI models fly drones?","url":"https://www.anthropic.com/research/project-pilot"},"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":"d9f71c58c11fbed0261bd97752ffefeeba10ede69ed0e5f6cab99bb7f632048e","title":"Anthropic and Andon Labs report on AI drone capabilities and oversight","argument":"On July 24, 2026, Anthropic and Andon Labs report on tests of AI models controlling a drone to find and follow a person in an office. Their benchmark, Drone-Bench, breaks the task into subtasks, including reconstructing the space, locating and navigating the drone, detecting a person, and following them. The best-performing model exceeded the human-AI team baseline on every task except reconstruction; in a real-drone demonstration, it followed the person better than the reference algorithm but could not navigate between rooms. The authors argue that this gap matters because increasingly capable systems could make human oversight seem like a cost rather than a safeguard, despite risks to physical security and privacy. They say decisions about oversight should be deliberate, and that investment in AI alignment, governance, and safety should increase with capability. The authors also stress that these results do not establish operational capability: tests used slow-moving drones in one office floorplan and did not cover outdoor settings or large crowds. Whether models can reliably pass the baseline while operating essentially autonomously remains an open question.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 29, 2026; originally published July 24, 2026. The evaluation used slow-moving drones in one office floorplan with a limited number of people; Andon Labs did not test outdoors in large crowds. Anthropic says it was not given access to Drone-Bench; Andon Labs ran the evaluations reported here.","speaker":{"name":"Anthropic","roleAtSource":"Publisher"},"source":{"title":"Project Pilot: Can AI models fly drones?","publisher":"Anthropic","format":"Article","url":"https://www.anthropic.com/research/project-pilot","publishedAt":"2026-07-24T00:00:00.000Z","dateBasis":"Date stated on the publisher’s page."},"evidence":[{"text":"Anthropic and Andon Labs report evaluating models on a drone-based locate-and-follow task and introduce Drone-Bench.","excerpt":"we developed a new series of demonstrations and evaluations that assess AI models’ ability to use a flying drone to autonomously perform a simple locate-and-follow task","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"Andon Labs created Drone-Bench, with Anthropic consulted, and ran the reported evaluations.","excerpt":"Drone-Bench is a benchmark created by Andon Labs (in consultation with Anthropic) to test if AI agents are capable of controlling a drone for surveillance tasks.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The best-performing model exceeded the baseline on every task except reconstruction, but could not autonomously navigate between rooms in the real-drone demonstration.","excerpt":"The best performing model was Claude Fable 5, which brings the frontier past the baseline on all tasks except reconstruction.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors argue that pressure to reduce oversight may rise when models pass capability and reliability thresholds, and that oversight decisions should be deliberate.","excerpt":"there will be real pressure to treat human oversight as a cost rather than a safeguard. That is exactly why these decisions must be made deliberately","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors say the evaluation does not establish operational capability and identify limits in its test setting.","excerpt":"the drones are moving at slow speeds, we only tested in one office floorplan with a limited number of people, and Andon did not test outdoors in large crowds","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"},{"text":"The authors argue that investment in AI alignment, governance, and safety should increase with the scale of capabilities.","excerpt":"the requirements for investing in AI alignment, governance, and safety increase with the scale of capabilities.","url":"https://www.anthropic.com/research/project-pilot","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.anthropic.com/research/project-pilot","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Can models operating essentially autonomously reliably pass the human-AI team baseline?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Anthropic and Andon Labs report on AI drone capabilities and oversight","summary":"On July 24, 2026, Anthropic and Andon Labs report on tests of AI models controlling a drone to find and follow a person in an office. Their benchmark, Drone-Bench, breaks the task into subtasks, including reconstructing the space, locating and navigating the drone, detecting a person, and following them. The best-performing model exceeded the human-AI team baseline on every task except reconstruction; in a real-drone demonstration, it followed the person better than the reference algorithm but could not navigate between rooms. The authors argue that this gap matters because increasingly capable systems could make human oversight seem like a cost rather than a safeguard, despite risks to physical security and privacy. They say decisions about oversight should be deliberate, and that investment in AI alignment, governance, and safety should increase with capability. The authors also stress that these results do not establish operational capability: tests used slow-moving drones in one office floorplan and did not cover outdoor settings or large crowds. Whether models can reliably pass the baseline while operating essentially autonomously remains an open question.","question":"Can models operating essentially autonomously reliably pass the human-AI team baseline?","claims":[{"text":"Anthropic and Andon Labs report evaluating models on a drone-based locate-and-follow task and introduce Drone-Bench.","excerpt":"we developed a new series of demonstrations and evaluations that assess AI models’ ability to use a flying drone to autonomously perform a simple locate-and-follow task","support":"source-stated"},{"text":"Andon Labs created Drone-Bench, with Anthropic consulted, and ran the reported evaluations.","excerpt":"Drone-Bench is a benchmark created by Andon Labs (in consultation with Anthropic) to test if AI agents are capable of controlling a drone for surveillance tasks.","support":"source-stated"},{"text":"The best-performing model exceeded the baseline on every task except reconstruction, but could not autonomously navigate between rooms in the real-drone demonstration.","excerpt":"The best performing model was Claude Fable 5, which brings the frontier past the baseline on all tasks except reconstruction.","support":"source-stated"},{"text":"The authors argue that pressure to reduce oversight may rise when models pass capability and reliability thresholds, and that oversight decisions should be deliberate.","excerpt":"there will be real pressure to treat human oversight as a cost rather than a safeguard. That is exactly why these decisions must be made deliberately","support":"source-stated"},{"text":"The authors say the evaluation does not establish operational capability and identify limits in its test setting.","excerpt":"the drones are moving at slow speeds, we only tested in one office floorplan with a limited number of people, and Andon did not test outdoors in large crowds","support":"source-stated"},{"text":"The authors argue that investment in AI alignment, governance, and safety should increase with the scale of capabilities.","excerpt":"the requirements for investing in AI alignment, governance, and safety increase with the scale of capabilities.","support":"source-stated"}],"limitations":["Text as retrieved on October 29, 2026; originally published July 24, 2026.","The evaluation used slow-moving drones in one office floorplan with a limited number of people; Andon Labs did not test outdoors in large crowds.","Anthropic says it was not given access to Drone-Bench; Andon Labs ran the evaluations reported here."],"mediaId":null,"correctionReason":null,"id":"d9f71c58c11fbed0261bd97752ffefeeba10ede69ed0e5f6cab99bb7f632048e","publicationId":"5dcd0673-488b-826a-0be5-40e8720d9e01","version":1,"sourceUrl":"https://www.anthropic.com/research/project-pilot","sourceTitle":"Project Pilot: Can AI models fly drones?","publisher":"Anthropic","publishedAt":"2026-07-24T00:00:00.000Z","originalPublished":{"value":"2026-07-24T00:00:00.000Z","precision":"instant","basis":"backfill:article_published"},"firstObservedAt":"2026-10-20T07:11:24.016Z","createdAt":"2026-11-06T07:38:24.468Z","historical":true,"entityIds":["anthropic"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.anthropic.com/research/project-pilot","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"b05bd850b63b952339576f2db1e0d72671c97b028a6e5dd67c7760114611a506","rubricVersion":"unrecorded","provider":"typesafe","model":"jev-1.13.0","reasonCodes":["bridge_admitted_by_model"],"confidence":0.62,"decidedAt":"2026-10-29T11:17:48.636Z"},"provenance":{"textHash":"3ab9772f682396ffe10025eac156ce610760f5c4df2ec119419668cf5a6f6bbd","observationId":"f3f57687d6da867f8afe9350df183e2aa1905b603f68a7a6a3fa21ed91482ad1","retrievedAt":"2026-10-29T11:19:31.016Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"3f15a897-fccb-a12d-a952-d4aa6c3646bc"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"1a4ee96c-bb26-94cd-837b-3e3ea93f0176"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-06T07:34:26.882Z","model":"anthropic/claude-opus-5.5"}}}],"audits":[{"publicationId":"5dcd0673-488b-826a-0be5-40e8720d9e01","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The page is credited to Anthropic and Andon Labs and published by Anthropic. The reading correctly says Andon Labs created Drone-Bench in consultation with Anthropic and ran the evaluations. It also keeps the oversight and governance arguments attributed to the authors rather than presenting them as established fact.","excerpt":"Anthropic has not been given access to Drone-Bench; Andon Labs ran the evaluations we report here."},{"dimension":"summary","status":"uncertain","explanation":"The summary is mostly faithful but has three gaps. (1) It says 'this gap matters because...', which ties the oversight argument to the reconstruction gap. The source ties pressure on oversight to models reliably passing the baseline. Smallest repair: 'the authors argue that once models pass this baseline, there may be pressure...'. (2) It omits a key reliability finding: Fable 5 reaches the baseline on average on only three of five tasks, so passing on every task but reconstruction was a one-off frontier result. (3) It does not explain plainly that the baseline came from Andon staff working with coding agents.","excerpt":"even Fable 5, the current frontier model, reaches the human baseline on average for only three of the five tasks"},{"dimension":"date","status":"uncertain","explanation":"July 24, 2026 matches the UTC day of originalPublished. There is no hindsight framing, and the reading does not rely on later passages. However, editedAfterPublication is true: the page was modified on 2026-10-23 per json-ld dateModified. The retrieved text cannot show which passages are original and which come from that later edit.","excerpt":null},{"dimension":"support","status":"uncertain","explanation":"Each listed claim has a matching verbatim excerpt. Two problems remain. First, the summary's causal link between 'this gap' and pressure on oversight is not stated by the source. Second, the page was edited after publication (dateModified 2026-10-23), so it cannot be confirmed that the cited passages, such as the Fable 5 results and the oversight argument, belong to the original July piece rather than the later edit.","excerpt":"Once models pass capability and reliability thresholds (such as the human-AI team baseline we used in this experiment), there will be real pressure to treat human oversight as a cost rather than a safeguard."},{"dimension":"media","status":"pass","explanation":"No media candidates were supplied, and the reading selects none. This is consistent with the evidence.","excerpt":null}],"createdAt":"2026-11-06T07:34:26.882Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}