EU AI Act Conformity Assessment Evidence Guide

SWT3 Witness Anchor Evidence for Regulation (EU) 2024/1689
Critical Assessor Notice: Boundaries of Cryptographic Evidence

SWT3 witness anchors prove that specific operational controls were active at a specific point in time. They do not replace the assessor's independent judgment, professional expertise, or regulatory authority. Assessors must verify that anchored evidence is sufficient, appropriate, and relevant to the specific assessment context. Conformity assessment remains the exclusive competence of the Notified Body per Article 43.

Version:
1.0
Date:
August 2026
Publisher:
Tenable Nova LLC
Audience:
Notified Body Assessors, Conformity Assessment Bodies
Regulation:
Regulation (EU) 2024/1689

Contents

  1. Purpose
  2. Scope and Enforcement Timeline
  3. Conformity Assessment Procedures
  4. High-Risk AI Evidence Mapping (Art. 9-15)
  5. GPAI Model Evidence Mapping (Art. 51-56)
  6. Limitations
  7. Evidence Collection Workflow
  8. Sample Assessment Findings
  9. Annex IV Documentation Requirements
  10. Harmonised Standards Landscape
  11. Penalty Tier Reference (Art. 99)
  12. Cross-References

1. Purpose

This guide maps SWT3 witness anchor evidence to the conformity assessment requirements of Regulation (EU) 2024/1689 (the EU Artificial Intelligence Act). It is written for Notified Body assessors performing conformity assessments under Article 43, and for conformity assessment bodies evaluating AI systems against the technical requirements in Chapters III and V.

For each applicable article, this guide identifies:

SWT3 provides cryptographic evidence of operational facts. It does not replace policy documentation, risk narratives, or organizational governance structures. This guide is explicit about those boundaries.

2. Scope and Enforcement Timeline

This guide covers two tracks within the EU AI Act:

TrackChapterArticlesEnforcement Date
High-Risk AI Systems (Annex III, standalone)Chapter III, Section 2Art. 9, 10, 11, 12, 13, 14, 152 December 2027 (Digital Omnibus deferral)
High-Risk AI Systems (Annex I, safety components)Chapter III, Section 2Art. 9, 10, 11, 12, 13, 14, 152 August 2028 (Digital Omnibus deferral)
GPAI ModelsChapter VArt. 51, 53, 55, 562 August 2025 (enforceable now)
GPAI obligations are already enforceable. Article 53 (general GPAI obligations) and Article 55 (systemic risk GPAI obligations) became enforceable on 2 August 2025. Providers of GPAI models with systemic risk must already comply with adversarial testing, incident reporting, and cybersecurity requirements.

Scope of application: High-risk AI systems as classified under Article 6 (systems listed in Annex III or used as safety components of products covered by Annex I). GPAI models as defined in Article 3(63), with systemic risk classification under Article 51.

3. Conformity Assessment Procedures

The EU AI Act defines two conformity assessment procedures for high-risk AI systems:

Article 43(1): Internal Control (Annex VI)

The provider performs a self-assessment against the requirements of Section 2 (Articles 9-15). SWT3 evidence serves as the provider's internal compliance evidence. The provider must demonstrate that their quality management system produces cryptographic proof of ongoing compliance.

How SWT3 fits: Anchor chains serve as the automatic documentation of operational compliance. The provider presents anchor history, compliance passports, and verification results as evidence of their internal control process.

Article 43(2): Conformity Assessment with NB Involvement (Annex VII)

A Notified Body evaluates the provider's quality management system and technical documentation. SWT3 evidence supports the NB's assessment by providing independently verifiable proof of runtime behavior.

How SWT3 fits: The NB independently verifies anchor fingerprints using standard SHA-256 (no proprietary tooling required). The NB reviews procedure coverage, verdict distribution, and clearing level appropriateness. The NB can spot-check any individual anchor by recomputing its fingerprint from the factor inputs.

No proprietary tooling required for verification. SWT3 fingerprints are computed from standard SHA-256 (RFC 6234). An assessor can verify any anchor using any SHA-256 implementation. The public verifier at sovereign.tenova.io/verify provides a convenient interface, but it is not required. See the SWT3 Protocol Specification v1.0 for the complete verification algorithm.

4. High-Risk AI Evidence Mapping

Article 9: Risk Management System Art. 99(2): 15M EUR / 3%

Regulatory Summary

Providers of high-risk AI systems shall establish, implement, document, and maintain a risk management system throughout the AI system's lifecycle. The system shall identify and analyse known and reasonably foreseeable risks, adopt suitable risk management measures, and ensure residual risks are communicated to deployers.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-MDL.1Model Identity VerificationArt. 9(4)(a)fa: hash present (1/0), fb: hash match (1/0), fc: method code
AI-GRD.1Guardrail EnforcementArt. 9(2)(a)fa: guardrails required, fb: guardrails active, fc: 0
AI-GRD.2Content Safety ClassificationArt. 9(4)(b)fa: categories checked, fb: passed (1/0), fc: 0
AI-FAIR.2Fairness Threshold CalibrationArt. 9(6)fa: threshold (scaled), fb: observed (scaled), fc: 0
AI-GOV.1Acceptable Use Policy AttestationArt. 9(1)fa: policy hash present (1/0), fb: enforced (1/0), fc: 0
AI-ACC.1Agent Access ControlArt. 9(4)(c)fa: scopes requested, fb: scopes granted, fc: 0
AI-SEC.1Adversarial Threat DetectionArt. 9(2)fa: tests run, fb: tests passed, fc: severity code

Sample Anchor

SWT3-E-VULTR-AI-AIGRD1-PASS-1774800000-a0aa7669ae6f Factors: factor_a=2 (2 guardrails required), factor_b=3 (3 guardrails active), factor_c=0 Verdict: PASS (active >= required) Interpretation: The AI system had 3 active guardrails at inference time, exceeding the required minimum of 2. Risk mitigation measures under Art. 9(2)(a) were in effect.

What an Assessor Should Verify

Partial SWT3 provides evidence that risk mitigation measures are operational at runtime. The risk management system documentation itself (risk identification, risk analysis, residual risk communication) is supplementary.

Supplementary Evidence Required Risk management system documentation (Art. 9(1)), risk identification and analysis records (Art. 9(2)(a)), residual risk communication to deployers (Art. 9(6)), testing and validation plans for risk mitigation measures.

Article 10: Data and Data Governance Art. 99(2): 15M EUR / 3%

Regulatory Summary

Training, validation, and testing data sets shall be subject to appropriate data governance and management practices. Data sets shall be relevant, sufficiently representative, and to the best extent possible free of errors and complete.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-DATA.1Training Data ProvenanceArt. 10(2)(a)fa: source count, fb: verified (1/0), fc: 0
AI-DATA.2Data Quality AssessmentArt. 10(2)(f)fa: quality score (scaled), fb: threshold (scaled), fc: 0
AI-DATA.3Training Data StatisticsArt. 10(3)fa: row count, fb: feature count, fc: class balance (scaled)
AI-DATA.4PII Lifecycle ManagementArt. 10(5)fa: record count, fb: pseudonymized (1/0), fc: method code
AI-FAIR.1Bias DetectionArt. 10(3)fa: threshold (scaled), fb: observed score (scaled), fc: 0
AI-GRD.3PII Redaction at InferenceArt. 10(2)(f)fa: patterns required, fb: patterns enforced, fc: 0

Sample Anchor

SWT3-E-VULTR-AI-AIDATA3-PASS-1774800036-f9ce447c6e19 Factors: factor_a=50000 (50,000 training rows), factor_b=128 (128 features), factor_c=850 (85.0% class balance, scaled by 10) Interpretation: Training data statistics witnessed. 50,000 rows across 128 features with 85% class balance. Assessor can verify the data governance practices referenced in the provider's technical documentation match these operational figures.

What an Assessor Should Verify

Partial SWT3 witnesses data quality metrics, bias scores, and PII lifecycle events. Training data governance policies, data provenance documentation, and representativeness analysis are supplementary evidence the provider must supply.

Supplementary Evidence Required Data governance and management practice documentation (Art. 10(2)), training data representativeness analysis (Art. 10(3)), data collection methodology, annotation guidelines, data sheet or data card.

Article 11: Technical Documentation Art. 99(2): 15M EUR / 3%

Regulatory Summary

Technical documentation shall be drawn up before the AI system is placed on the market or put into service, and shall be kept up to date. It shall demonstrate compliance with the requirements set out in Section 2.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-MDL.1Model Identity VerificationArt. 11(1)fa: hash present (1/0), fb: hash match (1/0), fc: 0
AI-MDL.2Model Version TrackingArt. 11(1)fa: version registered (1/0), fb: version match (1/0), fc: 0
AI-SBOM.1AI Software Bill of MaterialsArt. 11(1)fa: component count, fb: all identified (1/0), fc: format code
AI-CHR.1Agent Charter AttestationArt. 11(1)fa: charter hash present (1/0), fb: hash match (1/0), fc: 0

What an Assessor Should Verify

Partial SWT3 anchors provide automatic documentation of model identity, version, and component composition. The narrative technical documentation required by Annex IV (general description, design specifications, development methodology) is supplementary.

Supplementary Evidence Required Annex IV technical documentation (general description, design specifications, development methodology, risk management documentation, changes log, standards applied).

Article 12: Record-Keeping Art. 99(2): 15M EUR / 3%

Regulatory Summary

High-risk AI systems shall technically allow for the automatic recording of events (logs) over the lifetime of the system. Logging capabilities shall ensure a level of traceability of the AI system's functioning throughout its lifecycle.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-INF.1Inference ProvenanceArt. 12(1)fa: witnessed (1), fb: recorded (1), fc: 0
AI-INF.2Inference Latency MonitoringArt. 12(2)fa: threshold (ms), fb: observed (ms), fc: 0
AI-INF.3Inference Volume TrackingArt. 12(2)fa: expected volume, fb: observed volume, fc: 0
AI-LOG.1Structured Log AttestationArt. 12(1)fa: log entries, fb: schema-compliant (1/0), fc: format code
AI-TOOL.1Tool Call AccountabilityArt. 12(1)fa: tool invoked (1), fb: result code, fc: success (1/0)
AI-ID.1Agent Identity AttestationArt. 12(1)fa: identity present (1), fb: verified (1), fc: 0
AI-AUDIT.1Audit Log IntegrityArt. 12(1)fa: entries audited, fb: integrity verified (1/0), fc: 0

Sample Anchor

SWT3-E-VULTR-AI-AIINF1-PASS-1774800000-2e16e2fe92dd Factors: factor_a=1 (inference witnessed), factor_b=1 (recorded), factor_c=0 Interpretation: This anchor proves an AI inference event was automatically recorded with cryptographic integrity at the specified timestamp. The fingerprint is independently recomputable: SHA256("WITNESS:ENCLAVE_PROD:AI-INF.1:1:1:0:1774800000000")[:12]

What an Assessor Should Verify

Full SWT3 provides automatic recording of events with cryptographic integrity, timestamp, and independent verification. This directly satisfies the record-keeping requirement. The anchor chain itself is the log.

Supplementary Evidence Required None for the automatic recording requirement itself. The provider should document their retention policy (Art. 12(2): storage period appropriate for the intended purpose, at least 6 months).

Article 13: Transparency and Provision of Information Art. 99(2): 15M EUR / 3%

Regulatory Summary

High-risk AI systems shall be designed and developed in such a way as to ensure that their operation is sufficiently transparent to enable deployers to interpret the system's output and use it appropriately.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-EXPL.1Explainability AttestationArt. 13(1)fa: explanation present (1/0), fb: method code, fc: 0
AI-EXPL.2Confidence Score ReportingArt. 13(3)(b)(ii)fa: confidence threshold, fb: observed confidence, fc: 0
AI-TRANS.1Transparency NoticeArt. 13(3)(b)fa: notice displayed (1/0), fb: 1, fc: 0

What an Assessor Should Verify

Partial SWT3 witnesses explainability and confidence score generation. User-facing transparency notices, instructions for use, and deployer information packages are supplementary.

Supplementary Evidence Required Instructions for use document (Art. 13(3)), deployer information package, user-facing transparency notices and their placement, characteristics/limitations documentation (Art. 13(3)(b)).

Article 14: Human Oversight Art. 99(2): 15M EUR / 3%

Regulatory Summary

High-risk AI systems shall be designed and developed in such a way as to be effectively overseen by natural persons during the period in which they are in use. Human oversight measures shall aim to prevent or minimise risks, including through the ability to interrupt, override, or reverse AI system outputs.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-HITL.1Human Review AttestationArt. 14(1)fa: review required (1), fb: review completed (1), fc: 0
AI-HITL.2Override CapabilityArt. 14(4)(d)fa: override available (1), fb: override exercised (1/0), fc: 0
AI-HITL.3Reviewer Identity BindingArt. 14(1)fa: reviewers required, fb: reviewers present, fc: identity method code
AI-REV.1Output RevocationArt. 14(4)(d)fa: 1, fb: 1, fc: revocation reason code
AI-EMRG.1Emergency OverrideArt. 14(4)(e)fa: trigger type code, fb: response action code, fc: 0

Sample Anchor

SWT3-E-VULTR-AI-AIHITL3-PASS-1774800034-d7e3a52bd012 Factors: factor_a=2 (2 reviewers required), factor_b=2 (2 reviewers present), factor_c=2 (identity method: cryptographic binding) Interpretation: Four-eyes review completed with cryptographic identity binding. Two authorized reviewers confirmed the AI output before release. This provides evidence of Art. 14(1) human oversight with non-repudiation.

What an Assessor Should Verify

Partial SWT3 witnesses that human review was exercised, that override capability exists, and that revocation and emergency stop mechanisms function. However, Art. 14 also requires the system to be designed and developed to enable effective human oversight -- this is an architectural property that an external witness layer cannot fully attest. SWT3 proves oversight was exercised; supplementary evidence must demonstrate the system was designed to facilitate it.

Supplementary Evidence Required Documentation of which decisions require human review (risk-based classification), human oversight procedure documentation, training records for oversight personnel.

Article 15: Accuracy, Robustness, and Cybersecurity Art. 99(2): 15M EUR / 3%

Regulatory Summary

High-risk AI systems shall be designed and developed in such a way as to achieve an appropriate level of accuracy, robustness, and cybersecurity, and to perform consistently in those respects throughout their lifecycle.

SWT3 Procedures

ProcedureTitleArticle RefFactor Semantics
AI-DRIFT.1Model Drift DetectionArt. 15(1)fa: baseline score (scaled), fb: current score (scaled), fc: 0
AI-DRIFT.2Consequence-Mapped DriftArt. 15(1)fa: trigger type code, fb: response action code, fc: 0
AI-MDL.1Model Integrity VerificationArt. 15(3)fa: hash present (1/0), fb: hash match (1/0), fc: 0
AI-PERF.1Performance BenchmarkArt. 15(1)fa: benchmark score (scaled), fb: threshold (scaled), fc: benchmark type code
AI-ROBUST.1Robustness TestingArt. 15(3)fa: perturbation count, fb: survived count, fc: perturbation type code
AI-SEC.1Adversarial Threat DetectionArt. 15(4)fa: tests run, fb: tests passed, fc: severity code
AI-CYBER.1Cybersecurity PostureArt. 15(4)fa: controls assessed, fb: controls passing, fc: framework code

What an Assessor Should Verify

Partial SWT3 witnesses model integrity, drift detection, performance metrics, robustness testing, and cybersecurity posture at runtime. However, Art. 15 also requires appropriate accuracy levels to be achieved by design and declared in instructions for use -- these are design-time and documentation requirements that runtime monitoring alone cannot satisfy. SWT3 proves ongoing monitoring; supplementary evidence must demonstrate design-time accuracy targets and declared performance levels.

Supplementary Evidence Required Accuracy metrics methodology documentation, declared accuracy levels, error rates communication to deployers (Art. 15(2)), cybersecurity incident response plan.

5. GPAI Model Evidence Mapping (Art. 51-56)

This section maps SWT3 evidence to the obligations of providers of general-purpose AI models under Chapter V of the EU AI Act. These obligations apply to any organization that develops, trains, or substantially modifies a GPAI model and makes it available on the market.

Article 53: General Obligations for GPAI Providers

Regulatory Summary

Providers of GPAI models shall: (a) draw up and keep up to date technical documentation; (b) make available information and documentation to downstream providers integrating the model; (c) establish a policy to comply with copyright law; (d) make publicly available a sufficiently detailed summary of the training data.

SWT3 Procedures

ProcedureTitleArt. 53 RefFactor Semantics
AI-MDL.1Model Identity Verification53(1)(a)fa: hash present (1/0), fb: hash match (1/0), fc: 0
AI-MDL.2Model Version Tracking53(1)(a)fa: version registered (1/0), fb: version match (1/0), fc: 0
AI-SBOM.1AI Software Bill of Materials53(1)(a)fa: component count, fb: all identified (1/0), fc: format code
AI-CHR.1Agent Charter Attestation53(1)(a)fa: charter hash present (1/0), fb: hash match (1/0), fc: 0
AI-LIC.1License Compliance53(1)(c)fa: license type code, fb: compliant (1/0), fc: 0
AI-DEL.1Delegation Attestation53(1)(b)fa: deployer count, fb: required procedures, fc: delegation type code

Provider-Deployer Evidence Chain

Article 53(1)(b) requires GPAI providers to make information and documentation available to downstream providers. For organizations whose models are used by third parties -- whether through open-weight distribution, API access, cloud platform hosting, or embedded integration -- SWT3 provides a mechanism to demonstrate downstream compliance without accessing deployer infrastructure.

How it works:

  1. The model provider mints a delegation anchor (AI-DEL.1) recording the provider-deployer relationship.
  2. Downstream deployers integrate the SWT3 SDK (3 lines of code) and witness operations under their own tenant.
  3. Anchors flow to the deployer's own tenant, maintaining tenant isolation.
  4. The provider queries aggregated coverage metrics across deployer tenants: procedure coverage percentage, verdict distribution, clearing levels. The provider never sees raw deployer data.

This pattern is distribution-model-neutral:

Distribution ModelExampleDelegation Type
Open-weightModel weights distributed for local deploymentCode 0
API accessModel served via inference APICode 1
Cloud platformModel hosted on provider's cloud infrastructureCode 2
EmbeddedModel integrated into provider's applicationCode 3

Cross-provider interoperability: A deployer building on models from multiple providers uses a single SWT3 integration. The same anchors serve all provider relationships.

Partial SWT3 witnesses model identity, version tracking, SBOM, license compliance, and downstream deployer coverage. Technical documentation narrative, training data summaries, and copyright compliance policies are supplementary.

Supplementary Evidence Required Technical documentation per Annex XI (Art. 53(1)(a)), training data summary per Art. 53(1)(d), copyright compliance policy (Art. 53(1)(c)), downstream provider integration documentation (Art. 53(1)(b)).

Article 55: Obligations for Providers of GPAI with Systemic Risk

These obligations are enforceable now (since 2 August 2025). Providers of GPAI models classified with systemic risk under Art. 51 must already comply with these requirements.

Regulatory Summary

Providers of GPAI models with systemic risk shall: (a) perform model evaluation including adversarial testing; (b) assess and mitigate systemic risks; (c) report serious incidents to the AI Office; (d) ensure an adequate level of cybersecurity protection.

SWT3 Procedures

ProcedureTitleArt. 55 RefFactor Semantics
AI-REDTEAM.1Red Team Testing55(1)(a)fa: category code, fb: tests run, fc: vulnerabilities found
AI-ASSESS.1Champion-Challenger Assessment55(1)(a)fa: challenger score, fb: champion score, fc: threshold breached (0/1)
AI-INCIDENT.1Incident Report Attestation55(1)(c)fa: severity code, fb: incident type code, fc: reported (1/0)
AI-SEC.1Adversarial Threat Detection55(1)(d)fa: tests run, fb: tests passed, fc: severity code
AI-CYBER.1Cybersecurity Posture55(1)(d)fa: controls assessed, fb: controls passing, fc: framework code

What an Assessor Should Verify

Partial SWT3 provides evidence that adversarial testing, model evaluation, incident reporting, and cybersecurity measures are operational. Systemic risk assessment narratives, mitigation plans, and AI Office communications are supplementary.

Supplementary Evidence Required Systemic risk assessment and mitigation documentation (Art. 55(1)(b)), AI Office serious incident reports (Art. 55(1)(c)), state-of-the-art adversarial testing methodology documentation.

Article 56: Codes of Practice

Regulatory Summary

The AI Office shall encourage and facilitate the drawing up of codes of practice at Union level. GPAI model providers may rely on adherence to a code of practice to demonstrate compliance with Art. 53 and 55 obligations.

How SWT3 Supports Code-of-Practice Compliance

When a code of practice specifies operational requirements (e.g., "providers shall monitor model drift," "providers shall perform red-team testing quarterly"), SWT3 anchors provide cryptographic proof that those operational requirements were met. The anchor chain serves as a verifiable compliance log referenced by the code of practice.

Partial SWT3 provides evidence of operational adherence to code-of-practice requirements. The code of practice itself, governance documentation, and provider's declaration of adherence are supplementary.

6. Limitations

SWT3 provides cryptographic evidence of operational facts. It does not replace policy documentation, risk narratives, or organizational governance structures. The following items are explicitly outside the scope of SWT3 evidence:

An assessor encountering SWT3 evidence should evaluate it as one layer of a multi-layer evidence package. SWT3 anchors prove that controls were operational at runtime. Supplementary evidence proves that policies, procedures, and documentation exist and are adequate.

7. Evidence Collection Workflow

The following workflow is recommended for assessors evaluating SWT3 evidence as part of a conformity assessment:

StepActionVerification Method
1Request the provider's Compliance Passport or anchor exportJSON export, W3C Verifiable Credential, or HTML passport
2Verify anchor chain integrityIndependent fingerprint recomputation (SHA-256) for a random sample. Use the public verifier at sovereign.tenova.io/verify or compute locally.
3Review procedure coverage against required articlesMap provider's witnessed procedures to the article tables in Sections 4 and 5. Identify coverage gaps.
4Assess clearing level appropriatenessVerify that the clearing level does not impair traceability requirements (Art. 12). Level 3 removes model identity, which may limit Art. 12(2) compliance.
5Review supplementary evidenceFor each article marked PARTIAL, collect the supplementary evidence listed in the blue boxes above.
6Document findings with anchor referencesCite specific anchor tokens in conformity/non-conformity findings (see Section 8).

8. Sample Assessment Findings

Finding 1: Conformity (Art. 12 -- Record-Keeping) The provider demonstrates continuous automatic recording of inference events via AI-INF.1 witness anchors. A 30-day sample was reviewed: 47,231 anchors with 100% fingerprint recomputation success rate. Clearing level 1 preserves model identity for traceability. Anchor retention exceeds the 6-month minimum specified in Art. 12(2). Conformity with Art. 12(1) is established.
Finding 2: Non-Conformity (Art. 14 -- Human Oversight) AI-HITL.1 anchors are absent for high-risk decisions flagged in the provider's risk management documentation. The provider's risk assessment identifies credit scoring decisions as requiring human review, but no human review attestation anchors were found for the credit scoring model during the assessment period (1 June 2026 -- 30 June 2026). Non-conformity with Art. 14(1). Corrective action required: implement human review witnessing for all decision categories classified as high-risk in the provider's risk management system.
Finding 3: Observation (Art. 10 -- Data Governance) AI-DATA.3 anchors demonstrate training data statistics witnessing (50,000 rows, 128 features, 85% class balance). However, the training data governance policy documentation provided as supplementary evidence does not reference the SWT3 anchor chain as a data quality monitoring mechanism. Observation: The provider should align their data governance documentation with the operational evidence produced by the witnessing system to provide a complete evidence chain for Art. 10(2).
Finding 4: Conformity (Art. 55 -- GPAI Systemic Risk) The provider demonstrates adversarial testing via AI-REDTEAM.1 anchors (quarterly cadence, 4 test campaigns in the assessment period). AI-INCIDENT.1 anchors are present for 2 reported incidents with confirmed AI Office notification (factor_c=1). AI-CYBER.1 anchors show continuous cybersecurity posture monitoring. Conformity with Art. 55(1)(a), (c), and (d) is established. Note: systemic risk assessment narrative (Art. 55(1)(b)) was evaluated as supplementary evidence.

9. Annex IV Documentation Requirements

The following table maps Annex IV technical documentation sections to SWT3 procedures that provide supporting evidence:

Annex IV SectionRequirementSupporting SWT3 ProceduresCoverage
1General description of the AI systemAI-MDL.1, AI-MDL.2, AI-ID.1, AI-CHR.1Partial
2Detailed description of elements and development processAI-SBOM.1, AI-MDL.5, AI-MDL.6Partial
3Monitoring, functioning, and controlAI-DRIFT.1, AI-DRIFT.2, AI-PERF.1Partial
4Risk management systemAI-GRD.1, AI-VIO.1, AI-SAFE.1, AI-REACH.1Partial
5Intended purpose, foreseeable misuseAI-GOV.1, AI-CHR.1Partial
6Human oversight measuresAI-HITL.1, AI-HITL.2, AI-HITL.3, AI-REV.1Partial
7Accuracy, robustness, cybersecurityAI-DRIFT.1, AI-ROBUST.1, AI-SEC.1, AI-CYBER.1Partial
8Information about dataAI-DATA.1, AI-DATA.3, AI-FAIR.1Partial
9Pre-defined changesAI-MDL.2, AI-FREEZE.1Partial
10Logging capabilitiesAI-INF.1, AI-LOG.1, AI-AUDIT.1Full

Note: Annex IV sections marked PARTIAL require narrative documentation in addition to SWT3 operational evidence. SWT3 anchors supplement Annex IV but do not replace the narrative documentation requirement.

10. Harmonised Standards Landscape

The EU AI Act anticipates the development of harmonised European standards (hENs) to provide a presumption of conformity with the requirements in Chapter III. As of August 2026, the European Commission has issued standardisation requests to CEN and CENELEC (Standardisation Request M/593). Key standards in development include:

SWT3 evidence supports compliance with these standards by providing the operational evidence layer that management system documentation references. When harmonised standards are formally designated, SWT3 anchor chains can serve as the runtime evidence that demonstrates conformity with the technical requirements specified in the hEN.

Assessors should note that until hENs are formally designated in the Official Journal, compliance with ISO/IEC 42001 or other voluntary standards does not create a presumption of conformity with the EU AI Act. SWT3 evidence should be evaluated directly against the article requirements mapped in Sections 4 and 5.

Audit Portal

When working with a provider's audit portal share link, the EU AI Act Conformity section displays per-article compliance status with enforcement countdown, coverage badges, and exclusion management. The portal maps live evidence directly to the article requirements in Sections 4 and 5 of this guide. See the Assessor Onboarding Kit for a full portal walkthrough.

11. Penalty Tier Reference (Art. 99)

TierMaximum PenaltyApplies To
Art. 99(1)35,000,000 EUR or 7% of worldwide annual turnoverProhibited AI practices (Art. 5)
Art. 99(2)15,000,000 EUR or 3% of worldwide annual turnoverNon-compliance with Art. 9-15 requirements (high-risk AI)
Art. 99(3)7,500,000 EUR or 1.5% of worldwide annual turnoverIncorrect, incomplete, or misleading information to authorities

SWT3 evidence directly addresses Tier 2 (Art. 99(2)) obligations. Continuous witnessing provides a defense of due diligence: the provider can demonstrate that controls were operational and monitored throughout the system's lifecycle.

12. Cross-References