{"catalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","kind":"record","record":{"attributes":{"category":"administrative","domain":"AI Governance","type":"detective"},"canonicalUrl":"https://controlsmap.com/?v=1&node=uc%3AUC-AI-21","description":"Engage an independent evaluator at least quarterly to test each in-scope AI system against its defined risk categories: adversarial robustness and jailbreak resistance, harmful and out-of-scope outputs, agent-specific high-risk outputs, hallucination rates, and unsafe or unauthorized tool calls. Fix the test scope and pass thresholds in advance, track every finding to remediation and retest, and retain evaluator reports, methodologies, and remediation evidence.","details":{"control_category":"administrative","control_type":"detective","domain":"AI Governance","guidance":[],"members":[{"control_id":"B001","coverage":"full","framework":"aiuc-1","relationship":"superset_of"},{"control_id":"C010","coverage":"full","framework":"aiuc-1","relationship":"superset_of"},{"control_id":"C011","coverage":"full","framework":"aiuc-1","relationship":"superset_of"},{"control_id":"C012","coverage":"full","framework":"aiuc-1","relationship":"superset_of"},{"control_id":"D002","coverage":"full","framework":"aiuc-1","relationship":"superset_of"},{"control_id":"D004","coverage":"full","framework":"aiuc-1","relationship":"superset_of"}],"statement":"Engage an independent evaluator at least quarterly to test each in-scope AI system against its defined risk categories: adversarial robustness and jailbreak resistance, harmful and out-of-scope outputs, agent-specific high-risk outputs, hallucination rates, and unsafe or unauthorized tool calls. Fix the test scope and pass thresholds in advance, track every finding to remediation and retest, and retain evaluator reports, methodologies, and remediation evidence.","title":"Commission independent third-party AI evaluations on a quarterly cadence","unified_id":"UC-AI-21"},"id":"uc:UC-AI-21","mapUrl":"https://controlsmap.com/?v=1&node=uc%3AUC-AI-21","sourceIds":["aiuc-1"],"sourceUrl":null,"title":"UC-AI-21 — Commission independent third-party AI evaluations on a quarterly cadence","type":"unified"},"relationships":[{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:1df7d03ef432f33e2aa5d7ac1451629578a6aa10389545ab9db5a521f3bbf50a","properties":{"control_id":"C010","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-c010-b0a758b4.json","targetId":"ctrl:aiuc-1:C010","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:3725ab082d6da8f966bc8172699871ad9290c976ed4a2cb8b98b5400c81d7929","properties":{"control_id":"C012","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-c012-783c55df.json","targetId":"ctrl:aiuc-1:C012","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:38e45ec273a046a4eb2d0673c5ef8341389d4f59787559a2241ccfe927411245","properties":{"rationale":"Quarterly independent adversarial-robustness and jailbreak testing is the assurance that injection defenses actually hold.","strength":"primary"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/risk-ai-adversarial-poisoning-attacks-ea7df068.json","targetId":"risk:ai-adversarial-poisoning-attacks","type":"mitigates"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:4e40583b586e1bacd07058e5fd44e1e2d9267aed80d5f8be9fc976f65692227b","properties":{"rationale":"A recurring quarterly evaluation cadence catches degradation between releases that internal monitoring misses.","strength":"related"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/risk-ai-model-drift-monitoring-15329f38.json","targetId":"risk:ai-model-drift-monitoring","type":"mitigates"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:5972197934116a05943ce84de9076e56885db5794a5be749038b3e4728db3064","properties":{"rationale":"Independent hallucination-rate testing measures whether reliability controls are effective.","strength":"related"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/risk-ai-inaccurate-unreliable-output-cf10c86e.json","targetId":"risk:ai-inaccurate-unreliable-output","type":"mitigates"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:6a3d47be993481d1b17b86ce4d5f82bf4724655d6804992f80fcfb94036fb8e7","properties":{},"sourceDetailPath":"/data/v1/records/wf-c67-2d6bb2c2.json","sourceId":"wf:C67","targetDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","targetId":"uc:UC-AI-21","type":"operates"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:72b94ed91e964541cc87bf386c4b6994337ba2ac851d16d3f3454cfd9b8d8eb3","properties":{"control_id":"D002","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-d002-34554b38.json","targetId":"ctrl:aiuc-1:D002","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:7378e9eab62be8ba3d1a1ed5bd35f653d390c7165684821983a39c6e739cf851","properties":{"control_id":"C011","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-c011-7889db41.json","targetId":"ctrl:aiuc-1:C011","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:9a1a2dc56bb78b8ff0521c05adc3647c37d95cdf99641219b074b7ae7175865d","properties":{"rationale":"Adversarial-robustness testing includes extraction and over-disclosure probes against the endpoints.","strength":"related"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/risk-ai-endpoint-abuse-model-extraction-a3c97c52.json","targetId":"risk:ai-endpoint-abuse-model-extraction","type":"mitigates"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:b49441ad04e7bb2572f9ca20f5e87f347a22f75d23a7969d40c00117e61d6d09","properties":{"control_id":"D004","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-d004-21d4eeb4.json","targetId":"ctrl:aiuc-1:D004","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:e606dc337059883eca825fc58e55c1db8da105553e224d4723b18417ca8583ac","properties":{"control_id":"B001","coverage":"full","delta":null,"framework":"aiuc-1","provenance":{"defaultConfidence":"medium","defaultStatus":"active","direction":"canonical_to_source","mapper":"coworkcanvas-compliance-graph","note":"Each member is a documented relationship claim from the canonical unified control to a source control or guidance proposition. relationship: equal|superset_of (full) / intersects_with|subset_of (partial) / informs (guidance). confidence 'medium' = single-mapper, documented, not yet externally corroborated. source_version is the member framework's edition from the standard version register.","reviewDate":"2026-09-07"},"relationship":"superset_of","source_version":"July 15, 2026 release (quarterly update cadence)"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/ctrl-aiuc-1-b001-19a54ee3.json","targetId":"ctrl:aiuc-1:B001","type":"maps_to"},{"expectedCatalogRevision":"24028ffcfc2b295fa1b08ee6caa84b765f0731b321496bf4f548c49ad2177028","id":"rel:ee49bd21d5d9a068a75ee78a17c814da423e578682ac0adbb51894dd5349652c","properties":{"rationale":"Independent testing of tool calls surfaces unsafe or unauthorized agent actions before customers do.","strength":"related"},"sourceDetailPath":"/data/v1/records/uc-uc-ai-21-8afd47db.json","sourceId":"uc:UC-AI-21","targetDetailPath":"/data/v1/records/risk-ai-agent-unauthorized-actions-194415b8.json","targetId":"risk:ai-agent-unauthorized-actions","type":"mitigates"}],"schemaVersion":1}