{"schemaVersion":"nuwa-current-record-1","version":"2026-10-29.1","publication":"published","work":{"id":"loss-of-oversight","title":"Loss of Oversight: How AI systems may become harder to audit, monitor, and investigate","aliases":[],"authors":[],"primaryFunction":"evaluation","secondaryFunctions":[],"kind":"technical-report","sourceId":"loss-of-oversight-publication","attributionSourceId":"loss-of-oversight-publication","originalUrl":"https://www.aisi.gov.uk/research/loss-of-oversight-how-ai-systems-may-become-harder-to-audit-monitor-and-investigate","publicationDate":"2026-05-21","updatedOn":"2026-10-25","areas":[],"scopeNote":{"en":"Relevance is assigned to the research question and stated conditions of the work. A topic tag is not a risk rating.","zh":"风险关联依据研究问题及明确条件建立，主题标签不代表风险评级。"},"protocolIds":[],"contributions":[{"organizationId":"aisi","role":"publisher","sourceId":"loss-of-oversight-publication","locator":"Official publication masthead and article byline"}],"summary":{"en":"Combines expert interviews and literature analysis to examine the reliability of behavior, reasoning, internal and memory-based oversight signals.","zh":"结合专家访谈与文献分析，考察行为、推理、内部表征和记忆等监督信号的可靠性。"},"note":{"question":null,"method":null,"finding":null,"supports":null,"limits":null,"connections":null},"resources":[{"label":"Read the full paper","url":"https://cdn.prod.website-files.com/663bd486c5e4c81588db7a1d/6a0ed93f9b4a6a65994235d8_Loss_of_Oversight%20(7).pdf"}],"noteStatus":{"question":"limited","method":"limited","finding":"limited","supports":"limited","limits":"limited","connections":"limited"},"noteAvailability":{"en":"A substantive NUWA research note is not yet available. The repeated draft sections were removed pending review of the original methods, findings and limitations; use the original publication for the research account.","zh":"实质性的 NUWA 研究笔记尚未完成。重复草稿章节已移除，待核查原始方法、发现与限制；研究内容请以原始出版物为准。"},"noteSourceLocator":"No complete section-level method/results/limitations review completed for this note; cached original page and any chart audit retained as discovery evidence.","tags":["risk-monitoring"]},"reading":{"id":"loss-of-oversight","title":"Loss of Oversight: How AI systems may become harder to audit, monitor, and investigate","en":{"question":"Which present oversight affordances may erode as AI systems change?","design":"The official abstract describes 25 expert interviews plus literature review/internal analysis across behavior, CoT, internals, memory and honesty training.","findings":["The report identifies dependencies of current oversight and possible degradation paths.","Its recommendation is to track shifts, preserve affordances and develop fallback techniques."],"interpretation":"NUWA interpretation: this is structured foresight about oversight foundations, not a measured future failure rate.","limits":"Only the official abstract was read; interview recruitment, coding, disagreement and full argument details remain unaudited.","decisionRelevance":"Model recommendation: use the taxonomy for an oversight inventory while withholding quantitative predictions.","takeaway":"NUWA interpretation: this is structured foresight about oversight foundations, not a measured future failure rate.","connections":[],"scoring":null,"conditions":null,"statistics":null,"usage":"Model recommendation: use the taxonomy for an oversight inventory while withholding quantitative predictions."},"zh":{"question":"AI变化可能侵蚀哪些现有监督条件？","design":"官网摘要描述25次专家访谈，结合文献及内部分析，覆盖行为、思维链、内部表征、记忆与诚实训练。","findings":["报告梳理现有监督依赖及退化路径。","建议追踪变化、保留条件并开发备用方法。"],"interpretation":"NUWA 解读：这是监督基础的结构化前瞻，不是未来失败率测量。","limits":"仅阅读官网摘要，招募、编码、分歧及完整论证未审计。","decisionRelevance":"模型建议：用分类建立监督清单，暂不作数量预测。","takeaway":"NUWA 解读：这是监督基础的结构化前瞻，不是未来失败率测量。","connections":[],"scoring":null,"conditions":null,"statistics":null,"usage":"模型建议：用分类建立监督清单，暂不作数量预测。"},"assertionPolicy":{"question":"source-research-question","design":"source-finding","findings":"source-finding","interpretation":"nuwa-interpretation","limits":"source-limit-or-explicit-editorial-scope-limit","connections":"source-supported-relationship-only","decisionRelevance":"model-recommendation-not-source-conclusion"},"sourceRefs":[{"sourceId":"loss-of-oversight-publication","url":"https://www.aisi.gov.uk/research/loss-of-oversight-how-ai-systems-may-become-harder-to-audit-monitor-and-investigate","locator":"Official research page: Abstract","version":"Selected original resource https://www.aisi.gov.uk/research/loss-of-oversight-how-ai-systems-may-become-harder-to-audit-monitor-and-investigate; sha256:3e083ec0397c4173e626d57df659510646b7b62dcc9d693dfa92be0cd3145ada; archived source capture; exact capture time not independently established","artifactSha256":"3e083ec0397c4173e626d57df659510646b7b62dcc9d693dfa92be0cd3145ada"}],"publication":{"state":"published","publishedOn":"2026-10-29","authorizationReference":"user-authorized-global-index-publication-2026-10-29","reviewedCandidateId":"local-6efe31bc4a0c0f37"},"review":{"state":"published-note","sourceScope":"abstract","availableScope":null,"readingCoverage":"official-abstract-only; full original methods/results not audited","completePaperAudit":false,"independentEvidenceReview":"pending-human-review","bilingualEditorialReview":"pending-human-review","authorship":"model-assisted-research-note","preparedFingerprint":"b505662c3ea8978a32cce2e62d60e870724b16a4ff23ace684a3cb41951d80dd"}},"sources":[{"id":"loss-of-oversight-publication","title":"Loss of Oversight: How AI systems may become harder to audit, monitor, and investigate","url":"https://www.aisi.gov.uk/research/loss-of-oversight-how-ai-systems-may-become-harder-to-audit-monitor-and-investigate","version":"Source capture 2026-10-25","publicationDate":"2026-05-21","datePrecision":"day","type":"research-publication","verifiedOn":"2026-10-25","documentId":"https://www.aisi.gov.uk/research/loss-of-oversight-how-ai-systems-may-become-harder-to-audit-monitor-and-investigate"}]}