{ "schemaVersion": "2.0", "generatedAt": "2026-09-12T13:44:57.268Z", "compatibility": { "legacyMatterFieldsRetained": true, "issueDocumentsAreNotTimelineEvents": true }, "matter": { "id": "matter-29", "slug": "cnil-clearview-ai", "titleZh": "法国 CNIL 对 Clearview AI 处罚", "formalName": "CNIL Restricted Committee Decision SAN-2022-019 concerning Clearview AI", "aliases": [ "CNIL v. Clearview", "SAN-2022-019" ], "summary": "法国数据保护机构认定 Clearview 大规模抓取法国境内人员图像并建立生物识别数据库缺乏法律依据，处以 2,000 万欧元罚款并要求停止处理和删除数据。", "professionalSummary": "SAN-2022-019 依据 GDPR 处理合法性、数据主体权利与监管合作义务作出；命令还设置逾期按日罚款。", "whyItMatters": "它是欧洲监管机构针对通用人脸搜索训练数据库最严厉的公开行动之一。", "status": "under-order", "primaryTopic": "privacy-data-protection", "secondaryTopics": [ "public-law-human-rights" ], "aiRoles": [ "biometrics", "training-data" ], "jurisdiction": "法国·CNIL", "industries": [ "人脸识别", "数据经纪" ], "significance": { "legalNovelty": 28, "impactScope": 24, "institutionalLevel": 18, "crossBorder": 9, "sustainedCoverage": 14, "total": 93 }, "highProfile": true, "featured": true, "firstPublishedAt": "2026-07-15", "lastUpdatedAt": "2026-07-16", "proceedings": [ { "id": "matter-29:proceeding-1", "matterId": "matter-29", "formalName": "CNIL Restricted Committee Decision SAN-2022-019 concerning Clearview AI", "identifier": "SAN-2022-019", "identifiers": { "agencyFileNumber": "SAN-2022-019" }, "forum": "CNIL Restricted Committee", "jurisdiction": "法国·CNIL", "type": "administrative-enforcement", "stage": "final-order", "status": "under-order", "startedAt": "2022-10-17", "precedentialWeight": "not-applicable", "officialUrl": "https://www.cnil.fr/sites/default/files/atoms/files/deliberation_of_the_restricted_committee_no_san-2022-019_of_17_october_2022_concerning_clearview_ai.pdf" } ], "events": [ { "id": "matter-29:event-1", "matterId": "matter-29", "proceedingId": "matter-29:proceeding-1", "date": "2022-10-17", "type": "regulatory-action", "title": "CNIL 作出处罚决定", "summary": "CNIL 对 Clearview AI 处以 2,000 万欧元罚款，并命令停止无合法依据的收集处理和删除相关数据。", "documentIds": [ "matter-29:document-1" ], "claimIds": [ "matter-29:claim-1", "matter-29:claim-procedure", "matter-29:claim-event" ] } ], "documents": [ { "id": "matter-29:document-1", "familyId": "matter-29:family-1", "matterId": "matter-29", "eventId": "matter-29:event-1", "title": "Restricted Committee Deliberation No. SAN-2022-019", "originalTitle": "Restricted Committee Deliberation No. SAN-2022-019", "url": "https://www.cnil.fr/sites/default/files/atoms/files/deliberation_of_the_restricted_committee_no_san-2022-019_of_17_october_2022_concerning_clearview_ai.pdf", "kind": "official-document", "sourceId": "cnil", "publishedAt": "2022-10-17", "primaryOrSecondary": "primary", "relationship": "independent", "identifiers": { "agencyFileNumber": "SAN-2022-019" }, "language": "en" } ], "claims": [ { "id": "matter-29:claim-1", "matterId": "matter-29", "eventId": "matter-29:event-1", "text": "CNIL 对 Clearview AI 处以 2,000 万欧元罚款，并命令停止无合法依据的收集处理和删除相关数据。", "kind": "legal-holding", "status": "confirmed", "documentIds": [ "matter-29:document-1" ], "asOf": "2022-10-17" }, { "id": "matter-29:claim-procedure", "matterId": "matter-29", "eventId": "matter-29:event-1", "text": "SAN-2022-019 是本站记录的正式程序编号，程序由 CNIL Restricted Committee 处理。", "kind": "procedural-fact", "status": "confirmed", "documentIds": [ "matter-29:document-1" ], "asOf": "2022-10-17" }, { "id": "matter-29:claim-event", "matterId": "matter-29", "eventId": "matter-29:event-1", "text": "2022-10-17，案卷记录“CNIL 作出处罚决定”这一程序节点；该日期与事件由所引直接材料支持。", "kind": "procedural-fact", "status": "confirmed", "documentIds": [ "matter-29:document-1" ], "asOf": "2022-10-17" } ], "caseBackground": { "narrativeZh": "法国 CNIL 认定 Clearview AI 大规模抓取法国境内人员的网络图像，建立可供人脸搜索的生物识别数据库，却没有相应处理法律依据，并在数据主体权利及监管合作方面存在问题。CNIL 限制委员会在 SAN-2022-019 中作出处罚决定。", "partiesZh": [ "监管机构：法国国家信息与自由委员会（CNIL）", "被处罚方：Clearview AI" ], "aiSystemZh": "Clearview AI 人脸搜索与生物识别数据库", "challengedConductZh": "大规模抓取法国境内人员图像、生成生物识别模板并提供人脸搜索，同时未充分回应权利请求和监管要求。", "allegedHarmZh": "CNIL 认定相关人员的图像与生物识别数据在缺乏法律依据的情况下被处理。", "proceduralOriginZh": "CNIL 限制委员会依据 GDPR 进行行政执法，并于 2022 年 10 月 17 日作出 SAN-2022-019。", "claimIds": [ "matter-29:claim-1", "matter-29:claim-procedure", "matter-29:claim-event" ], "documentIds": [ "matter-29:document-1" ] }, "issues": [ { "id": "matter-29:issue-lawful-basis", "matterId": "matter-29", "titleZh": "抓取图像与生成生物识别模板的法律依据", "questionZh": "Clearview 是否具有大规模收集法国境内人员图像并处理为生物识别数据库的 GDPR 法律依据？", "contextZh": "公开可见的网络照片并不自动等于可以不受限制地用于生物识别训练与搜索。", "topic": "privacy-data-protection", "status": "resolved", "courtTreatmentZh": "CNIL 认定 Clearview 的收集和处理缺乏法律依据。", "claimIds": [ "matter-29:claim-1" ], "positions": [ { "id": "matter-29:issue-lawful-basis:position-1", "issueId": "matter-29:issue-lawful-basis", "sideZh": "CNIL", "positionZh": "认定大规模图像与生物识别处理没有合法基础。", "claimIds": [ "matter-29:claim-1" ], "documentIds": [ "matter-29:document-1" ] }, { "id": "matter-29:issue-lawful-basis:position-2", "issueId": "matter-29:issue-lawful-basis", "sideZh": "Clearview AI", "positionZh": "现有底库未收录其对法律依据的具体答辩。", "claimIds": [], "documentIds": [] } ] }, { "id": "matter-29:issue-data-subject-rights", "matterId": "matter-29", "titleZh": "数据主体权利的有效行使", "questionZh": "Clearview 是否充分保障法国数据主体查询、访问或删除其数据的权利？", "contextZh": "人脸数据库中的个人往往并非主动用户，因此权利请求机制是数据治理的核心保障。", "topic": "privacy-data-protection", "status": "open", "claimIds": [], "positions": [ { "id": "matter-29:issue-data-subject-rights:position-1", "issueId": "matter-29:issue-data-subject-rights", "sideZh": "CNIL", "positionZh": "认定 Clearview 在数据主体权利履行方面违反 GDPR。", "claimIds": [], "documentIds": [] }, { "id": "matter-29:issue-data-subject-rights:position-2", "issueId": "matter-29:issue-data-subject-rights", "sideZh": "Clearview AI", "positionZh": "底库未记录其针对具体权利请求处理的答辩。", "claimIds": [], "documentIds": [] } ] }, { "id": "matter-29:issue-regulatory-cooperation", "matterId": "matter-29", "titleZh": "跨境企业的监管合作义务", "questionZh": "Clearview 是否履行了配合 CNIL 调查与执行要求的义务？", "contextZh": "跨境数据企业即使没有面向法国用户的传统账户关系，也可能受到法国监管调查。", "topic": "public-law-human-rights", "status": "open", "claimIds": [], "positions": [ { "id": "matter-29:issue-regulatory-cooperation:position-1", "issueId": "matter-29:issue-regulatory-cooperation", "sideZh": "CNIL", "positionZh": "认定 Clearview 未充分履行监管合作义务。", "claimIds": [], "documentIds": [] }, { "id": "matter-29:issue-regulatory-cooperation:position-2", "issueId": "matter-29:issue-regulatory-cooperation", "sideZh": "Clearview AI", "positionZh": "现有底库未记录其对合作义务认定的详细回应。", "claimIds": [], "documentIds": [] } ] } ], "decisions": [ { "id": "matter-29:decision-1", "matterId": "matter-29", "proceedingId": "matter-29:proceeding-1", "eventId": "matter-29:event-1", "date": "2022-10-17", "outcomeZh": "CNIL 对 Clearview AI 处以 2,000 万欧元罚款，并命令停止没有法律依据的相关处理和删除数据。", "holdingsZh": [ "大规模收集法国境内人员图像并建立生物识别数据库缺乏法律依据。", "Clearview 在数据主体权利和监管合作义务方面违反 GDPR。", "停止处理和删除相关数据是本案的核心纠正措施。" ], "reasoningZh": [ "网络图像可公开访问并不消除将其转化为敏感生物识别数据所需的 GDPR 法律基础。", "处理规模、数据敏感性以及权利救济和监管合作缺陷共同支持高额罚款和删除命令。" ], "reliefZh": "2,000 万欧元行政罚款；停止无合法依据的收集处理；删除相关数据；逾期履行还受按日罚款约束，底库未列明按日金额。", "scopeAndLimitsZh": "决定针对 Clearview 涉及法国境内人员数据的处理及其 GDPR 义务，不是对所有人脸识别技术的一概禁令。", "precedentialWeight": "not-applicable", "appealStatusZh": "SAN-2022-019 于 2022 年 10 月 17 日作出，事项处于命令履行状态；底库未记录上诉或撤销。", "documentIds": [ "matter-29:document-1" ], "claimIds": [ "matter-29:claim-1" ] } ], "issueDocumentLinks": [ { "id": "news-context-link-matter-29-matter-29:issue-lawful-basis-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-lawful-basis", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 2 个确定性争点词。", "viewpoint": "critical", "confidence": 0.9, "assignedBy": "rule" }, { "id": "news-context-link-matter-29-matter-29:issue-data-subject-rights-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-data-subject-rights", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" }, { "id": "issue-document-matter-29-matter-29:issue-regulatory-cooperation-knowledge-perpetual-line-up", "matterId": "matter-29", "issueId": "matter-29:issue-regulatory-cooperation", "documentId": "knowledge-perpetual-line-up", "relationship": "empirical-context", "relevanceSummaryZh": "以公开记录调查说明执法场景中数据库规模、观察名单、准确率和监督机制为何需要分别核验。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词；AI 作用分面只用于候选召回。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" }, { "id": "news-context-link-matter-29-matter-29:issue-regulatory-cooperation-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-regulatory-cooperation", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" } ] }, "issueDocuments": [ { "link": { "id": "news-context-link-matter-29-matter-29:issue-lawful-basis-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-lawful-basis", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 2 个确定性争点词。", "viewpoint": "critical", "confidence": 0.9, "assignedBy": "rule" }, "document": { "id": "news-context-facial-recognition-challenges", "familyId": "news-family-facial-recognition-challenges", "title": "人脸识别技术带来的法律与社会挑战", "originalTitle": "The Challenges of Facial Recognition Technologies", "url": "https://hai.stanford.edu/news/challenges-facial-recognition-technologies", "kind": "factual-report", "sourceId": "stanford-hai", "publishedAt": "2020-04-23", "author": "Shana Lynch", "primaryOrSecondary": "secondary", "relationship": "independent", "identifiers": {}, "language": "en", "metadata": { "abstractZh": "报道以 Marietje Schaake 的研究和政策观察为主线，梳理人脸识别在偏差、无感收集、数据所有权、执法监控与跨场景滥用方面的问题。", "venue": "Stanford HAI News", "publisher": "Stanford Institute for Human-Centered Artificial Intelligence", "publicationState": "published", "peerReviewed": false, "paywalled": false }, "editorialSummaryZh": "这篇机构报道把商业认证、公共空间追踪与执法识别放在同一风险图景中，适合为 Clearview、BIPA 和中国人脸识别案件提供背景，但不评价具体算法的准确率。", "legalTopics": [ "privacy-data-protection", "discrimination", "public-law-human-rights" ], "aiRoles": [ "biometrics", "automated-decision" ] } }, { "link": { "id": "news-context-link-matter-29-matter-29:issue-data-subject-rights-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-data-subject-rights", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" }, "document": { "id": "news-context-facial-recognition-challenges", "familyId": "news-family-facial-recognition-challenges", "title": "人脸识别技术带来的法律与社会挑战", "originalTitle": "The Challenges of Facial Recognition Technologies", "url": "https://hai.stanford.edu/news/challenges-facial-recognition-technologies", "kind": "factual-report", "sourceId": "stanford-hai", "publishedAt": "2020-04-23", "author": "Shana Lynch", "primaryOrSecondary": "secondary", "relationship": "independent", "identifiers": {}, "language": "en", "metadata": { "abstractZh": "报道以 Marietje Schaake 的研究和政策观察为主线，梳理人脸识别在偏差、无感收集、数据所有权、执法监控与跨场景滥用方面的问题。", "venue": "Stanford HAI News", "publisher": "Stanford Institute for Human-Centered Artificial Intelligence", "publicationState": "published", "peerReviewed": false, "paywalled": false }, "editorialSummaryZh": "这篇机构报道把商业认证、公共空间追踪与执法识别放在同一风险图景中，适合为 Clearview、BIPA 和中国人脸识别案件提供背景，但不评价具体算法的准确率。", "legalTopics": [ "privacy-data-protection", "discrimination", "public-law-human-rights" ], "aiRoles": [ "biometrics", "automated-decision" ] } }, { "link": { "id": "issue-document-matter-29-matter-29:issue-regulatory-cooperation-knowledge-perpetual-line-up", "matterId": "matter-29", "issueId": "matter-29:issue-regulatory-cooperation", "documentId": "knowledge-perpetual-line-up", "relationship": "empirical-context", "relevanceSummaryZh": "以公开记录调查说明执法场景中数据库规模、观察名单、准确率和监督机制为何需要分别核验。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词；AI 作用分面只用于候选召回。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" }, "document": { "id": "knowledge-perpetual-line-up", "familyId": "knowledge-family-perpetual-line-up", "title": "永久列队：美国未受规制的警务人脸识别", "originalTitle": "The Perpetual Line-Up: Unregulated Police Face Recognition in America", "url": "https://www.law.georgetown.edu/privacy-technology-center/research/the-perpetual-line-up/", "kind": "academic-research", "sourceId": "georgetown-center-privacy-technology", "publishedAt": "2016-10-18", "author": "Clare Garvie; Alvaro M. Bedoya; Jonathan Frankle", "primaryOrSecondary": "secondary", "relationship": "independent", "identifiers": {}, "language": "en", "metadata": { "abstractZh": "依据一百余项信息公开请求和访谈调查警务人脸识别数据库的规模、准确率测试、观察名单、透明度和监督缺口，并把技术使用放回执法权力与程序保障中评价。", "venue": "Georgetown Law Center on Privacy & Technology Research Report", "publisher": "Georgetown University Law Center", "publicationState": "published", "peerReviewed": false, "paywalled": false }, "stableIdentifier": { "kind": "repository-record", "value": "georgetown-cpt:perpetual-line-up:2016" }, "legalTopics": [ "privacy-data-protection", "public-law-human-rights", "discrimination" ], "aiRoles": [ "biometrics", "automated-decision" ] } }, { "link": { "id": "news-context-link-matter-29-matter-29:issue-regulatory-cooperation-news-context-facial-recognition-challenges", "matterId": "matter-29", "issueId": "matter-29:issue-regulatory-cooperation", "documentId": "news-context-facial-recognition-challenges", "relationship": "news-context", "relevanceSummaryZh": "为生物识别案件提供无感收集、执法使用、偏差和跨场景滥用的社会与政策背景。 本次规则匹配仅使用本争点文本及其引用主张，命中 1 个确定性争点词。", "viewpoint": "critical", "confidence": 0.8500000000000001, "assignedBy": "rule" }, "document": { "id": "news-context-facial-recognition-challenges", "familyId": "news-family-facial-recognition-challenges", "title": "人脸识别技术带来的法律与社会挑战", "originalTitle": "The Challenges of Facial Recognition Technologies", "url": "https://hai.stanford.edu/news/challenges-facial-recognition-technologies", "kind": "factual-report", "sourceId": "stanford-hai", "publishedAt": "2020-04-23", "author": "Shana Lynch", "primaryOrSecondary": "secondary", "relationship": "independent", "identifiers": {}, "language": "en", "metadata": { "abstractZh": "报道以 Marietje Schaake 的研究和政策观察为主线，梳理人脸识别在偏差、无感收集、数据所有权、执法监控与跨场景滥用方面的问题。", "venue": "Stanford HAI News", "publisher": "Stanford Institute for Human-Centered Artificial Intelligence", "publicationState": "published", "peerReviewed": false, "paywalled": false }, "editorialSummaryZh": "这篇机构报道把商业认证、公共空间追踪与执法识别放在同一风险图景中，适合为 Clearview、BIPA 和中国人脸识别案件提供背景，但不评价具体算法的准确率。", "legalTopics": [ "privacy-data-protection", "discrimination", "public-law-human-rights" ], "aiRoles": [ "biometrics", "automated-decision" ] } } ], "sources": [ { "id": "cnil", "slug": "cnil", "name": "Commission nationale de l'informatique et des libertés", "homepage": "https://www.cnil.fr/", "type": "regulator", "grade": "A", "primaryOrSecondary": "primary", "score": { "traceability": 25, "corrections": 18, "ownershipTransparency": 20, "expertise": 20, "historicalAccuracy": 15, "total": 98 }, "rationale": "法院、政府或监管机构自有发布渠道；程序状态以其正式文书与登记信息为准。", "methodologyVersion": "1.0", "reviewedAt": "2026-07-15" }, { "id": "stanford-hai", "slug": "stanford-hai", "name": "Stanford Institute for Human-Centered Artificial Intelligence", "homepage": "https://hai.stanford.edu/", "type": "academic", "primaryOrSecondary": "mixed", "rationale": "大学研究机构发布研究、政策简报与新闻；按具体作者和文类区分一手研究与机构评论。", "grade": "A", "score": { "traceability": 23, "corrections": 16, "ownershipTransparency": 19, "expertise": 20, "historicalAccuracy": 12, "total": 90 }, "methodologyVersion": "1.0", "reviewedAt": "2026-07-16", "ingestion": { "feedEnabled": false, "feedUrl": null, "verification": "not-verified", "allowedUse": "discovery-metadata-and-deep-links" } }, { "id": "georgetown-center-privacy-technology", "slug": "georgetown-center-privacy-technology", "name": "Georgetown Law Center on Privacy & Technology", "homepage": "https://www.law.georgetown.edu/privacy-technology-center/", "type": "academic", "primaryOrSecondary": "secondary", "rationale": "大学法学院研究中心，调查报告公开作者、方法、原始信息请求范围及稳定项目页。", "grade": "A", "score": { "traceability": 25, "corrections": 18, "ownershipTransparency": 20, "expertise": 20, "historicalAccuracy": 14, "total": 97 }, "methodologyVersion": "1.0", "reviewedAt": "2026-07-16" } ], "canonicalUrl": "/cases/cnil-clearview-ai" }