Pre-conformity mock assessment
We run the Art. 43 conformity assessment route of a high-risk AI system as a dry run under PR-43-MOCK v1.0: internal control under Annex VI, or quality management system and technical documentation assessment under Annex VII. Each annex point is checked against the evidence for Art. 9 to Art. 17, and a sample of the tests is re-run. The client receives a dry-run review report, a findings register, an evidence-pack index and a remediation plan. Annex III systems meet the route from the Annex III application date Deferred. No certificate follows the dry run.
Deliverables
The set is designed to answer the question an assessor, a notified body or a market surveillance authority puts to the provider: which annex point is met by which document, which test result stands behind which declared value, and which points have no evidence. Each type is specified with a synthetic sample on the deliverables page: the dry-run review follows the assessment memo specification, the findings register and the remediation plan follow the findings register specification, and the evidence-pack index follows the evidence pack specification.
| Deliverable | Type | Format | Designed to satisfy |
|---|---|---|---|
| Dry-run review of the Annex VI or Annex VII route as an assessor would conduct it | Assessment memo | PDF with Markdown source | |
| Findings register with severity and article reference | Findings register | XLSX; CSV; PDF summary | |
| Evidence-pack index with gaps marked | Evidence pack | Folder with index (PDF, Markdown); ZIP with hash manifest | |
| Remediation plan | Findings register | XLSX; CSV; PDF summary |
Who this is for
- Annex III providers
- Annex I manufacturers
A provider of an Annex III system, for example recruitment screening, credit scoring, education admission or emergency-service triage, carries the full Chapter III Section 2 set from risk management to the quality management system, and most of those requirements have no numeric threshold in the Act; the provider sets and justifies each threshold, and an assessor checks the justification and the evidence behind it. A manufacturer whose AI is a safety component of an Annex I Section A product already runs a technical file, a quality system and a notified body relationship under the sectoral law, and the AI Act requirements are assessed inside that route, so the task is to add AI-specific evidence to the existing file without duplicating what the sectoral law already covers. In both cases the dry run shows, before the real assessment, which points hold and which do not.
Two entries set the urgency: the Annex III application date, from which the Art. 43 route applies to systems classified under Art. 6(2), and the Annex I date, from which it applies to products and safety components classified under Art. 6(1).
- DeferredAnnex III standalone high-risk obligations
- DeferredAnnex I product-embedded high-risk obligations
Scope
The engagement covers three groups of elements of PR-43-MOCK v1.0; the limits of the dry run are stated on what we cannot do. First, the route rules in Art. 43(1) to 43(4): an Annex III point 1 system goes to Annex VII unless the provider applied harmonised standards or common specifications in full, an Annex III points 2 to 8 system goes to internal control under Annex VI, an Annex I Section A system follows its sectoral conformity procedure with Annex VII points 4.3 to 4.5 inside it, and a substantial modification after placing on the market starts a new assessment. No harmonised standard is cited in the Official Journal as of the anchor date, so the Art. 43(1) condition that would open Annex VI to an Annex III point 1 system is not met, and no presumption of conformity under Art. 40(1) is available to any system; the standards page tracks the count. Second, each point of Annex VI or Annex VII as an assessor would check it: the quality management system against Art. 17, the technical documentation against Chapter III Section 2, the consistency between the documented process and the actual process, data-set access, surveillance and change notification. Third, the presence and content of the three items that follow a positive assessment: the EU declaration of conformity under Art. 47 with Annex V, the CE marking plan under Art. 48 and the registration record under Art. 49 with Annex VIII.
The quality management system check uses the Art. 17(1) list as its checklist. prEN 18286Published supplies the clause structure that PR-17-QMS follows, and ISO/IEC 42001:2023Published supplies the management-system structure at the next hierarchy level; neither is cited in the Official Journal, so both are evidence sources and neither creates a presumption. Within the second group the protocol samples the testable elements of the Art. 9, Art. 10, Art. 11, Art. 12, Art. 13, Art. 14, Art. 15 and Art. 17 records and re-runs the sample with the protocol named on each obligation page. A passing sample shows that the sampled evidence reproduces; the complete evidence for a record comes from that record's own protocol and service, for example Technical documentation and evidence-pack preparation for Art. 11 and Art. 13.
The engagement does not cover a system the provider has concluded is not high-risk under Art. 6(3); PR-CLS covers that determination, and the registration duty that follows such a conclusion is an unsettled point recorded in the Art. 9 obligation record. Machinery under Regulation (EU) 2023/1230 moved to Annex I Section B In force; for Section B products Art. 2(2) limits the Act to Art. 6(1), Art. 102 to Art. 109 and Art. 112, so the sectoral procedure carries the requirements and the dry run does not apply to them.
Where the sample draws the special-category data element of the Art. 10 record, the re-test under PR-10-BIAS checks the legal-basis record against the working reading of the Art. 4a necessity threshold, and the findings register states which reading the record meets.
The dry run is not a conformity assessment under Art. 43, and its conclusion is not a certificate, not a Union technical documentation assessment certificate under Annex VII point 4.6, and not a decision that binds a notified body or a market surveillance authority. A change since the last assessment is classified in the modification register under Art. 43(4); where a change is substantial, the Regression testing after substantial modification service continues from that register.
What we test
The table lists the testable elements of all eight obligation records this service covers, with the protocol that tests each element in full and the artifacts it produces for the record; the dry run samples from this table, and the table is long because the sample can draw from any row.
| Obligation | Testable element | Protocol | Artifacts |
|---|---|---|---|
| Art. 9 | The system performs consistently for its intended purpose across the operating conditions recorded in the risk management file, with a result per condition (Art. 9(6)). | PR-17-QMS v1.0 |
|
| Each test in the risk management file runs against a prior defined metric and probabilistic threshold, and the result either meets the threshold or does not (Art. 9(8)). | PR-17-QMS v1.0 | ||
| The residual risk of each hazard after the adopted measures is measured and compared with the acceptance criterion recorded for that hazard, and the overall residual risk is compared with the overall criterion (Art. 9(5)). | PR-17-QMS v1.0 | ||
| Each technical risk management measure adopted under Art. 9(2)(d), such as an input check, a filter, a fallback or a guardrail, activates on the hazard scenarios it addresses and produces the recorded effect (Art. 9(5)(a) and 9(5)(b)). | PR-17-QMS v1.0 | ||
| The system's behaviour under each reasonably foreseeable misuse condition recorded under Art. 9(2)(b) is observed and recorded against the expected behaviour. | PR-17-QMS v1.0 | ||
| Where the Art. 9(9) consideration identifies a likely adverse impact on persons under the age of 18 or other vulnerable groups, the test set contains cases that represent those groups and the results are reported for them. | PR-17-QMS v1.0 | ||
| Art. 10 | Representativeness of the training, validation and testing data sets against the persons or groups of persons the system is intended to be used for, measured as subgroup coverage against the declared intended purpose (Art. 10(3)) | PR-10-BIAS v1.0 |
|
| Error rate and completeness of the data sets, measured by label-error sampling and missing-value profiling against the targets the provider declares (Art. 10(3)) | PR-10-BIAS v1.0 | ||
| Statistical properties of the data sets, compared with the documented assumptions about what the data measure and represent (Art. 10(2)(d) and 10(3)) | PR-10-BIAS v1.0 | ||
| Coverage of the geographical, contextual, behavioural or functional setting of intended use in the data sets, checked against the setting declared in the intended purpose (Art. 10(4)) | PR-10-BIAS v1.0 | ||
| Bias examination result: system performance and outcomes disaggregated by protected and context-relevant subgroup, with each metric defined (Art. 10(2)(f)) | PR-10-BIAS v1.0 | ||
| Effect of the bias mitigation measures, verified by re-running the disaggregated evaluation after mitigation (Art. 10(2)(g)) | PR-10-BIAS v1.0 | ||
| Feedback-loop check where data outputs influence inputs for future operations, comparing subgroup distributions across operation cycles (Art. 10(2)(f)) | PR-10-BIAS v1.0 | ||
| Special-category data handling check: pseudonymisation and re-use limits present, access restricted to authorised persons, no transfer to other parties, observed in the data pipeline (Art. 4a(1)(b), (c) and (d)) | PR-10-BIAS v1.0 | ||
| Art. 11 | The accuracy, robustness and other metric values declared in the documentation match the values in the dated and signed test reports it contains, for the same system version and the same test conditions (Annex IV section 2(g)) | PR-11-DOC v1.0 |
|
| The degrees of accuracy stated for specific persons or groups of persons and the overall expected level of accuracy match the disaggregated and aggregate results in the test reports (Annex IV section 3) | PR-11-DOC v1.0 | ||
| The system version and the software or firmware versions stated in the general description match the deployed artefact and the versions named in the test reports (Annex IV section 1(a) and 1(c)) | PR-11-DOC v1.0 | ||
| The architecture description matches the system as built: the listed components, their interfaces and the integrations with external hardware, software and other AI systems are present and no deployed component is missing from the description (Annex IV section 2(c) and 1(b)) | PR-11-DOC v1.0 | ||
| The intended purpose and the conditions of use in the general description match the instructions for use and the promotional or sales materials (Annex IV section 1(a) and 1(h)) | PR-11-DOC v1.0 | ||
| The datasheets match the data sets actually used: provenance, scope, main characteristics, labelling procedure and cleaning methodology as recorded in the data pipeline (Annex IV section 2(d)) | PR-11-DOC v1.0 | ||
| The human oversight measures and the output interpretation measures described in the documentation are present in the system and in the instructions for use (Annex IV section 2(e) and section 3) | PR-11-DOC v1.0 | ||
| The cybersecurity measures described in the documentation are present in the deployed configuration (Annex IV section 2(h)) | PR-11-DOC v1.0 | ||
| Every test log and test report in the documentation carries a date and the signature of a responsible person, and the set covers each pre-determined change described in section 2(f) (Annex IV section 2(g)) | PR-11-DOC v1.0 | ||
| The lifecycle change record matches the version history in the general description and the versions covered by test reports (Annex IV section 6 with section 1(a) and 1(c)) | PR-11-DOC v1.0 | ||
| Each entry in the standards and specifications list maps to the requirement it addresses, and each Chapter III Section 2 requirement without an applied harmonised standard maps to a described solution (Annex IV section 7) | PR-11-DOC v1.0 | ||
| The copy of the EU declaration of conformity names the same system, version and provider as the general description (Annex IV section 8 with section 1(a)) | PR-11-DOC v1.0 | ||
| Art. 12 | Events are recorded automatically by the system, without an operator action, during normal operation (Art. 12(1)) | PR-12-LOG v1.0 |
|
| Logging continues to function, with the same event set and format, after a model, software or configuration update over the lifetime of the system (Art. 12(1)) | PR-12-LOG v1.0 | ||
| A sampled decision can be reconstructed from the logs alone: the input, the model version, the configuration in force, the output and the timestamp of the event (Art. 12(2)) | PR-12-LOG v1.0 | ||
| Each risk indicator named in the risk management file under Art. 9 has a log event that records it, and the event fires when the indicator condition is met (Art. 12(2)(a)) | PR-12-LOG v1.0 | ||
| A model version change, a threshold change or a configuration change after placing on the market produces a log event that identifies it as a candidate substantial modification (Art. 12(2)(a)) | PR-12-LOG v1.0 | ||
| Each data point the post-market monitoring plan under Art. 72 consumes is present in the logs with the field definition the plan expects (Art. 12(2)(b)) | PR-12-LOG v1.0 | ||
| The logs the deployer needs for monitoring operation under Art. 26(5) are readable and exportable by the deployer through the mechanism described in the instructions for use (Art. 12(2)(c)) | PR-12-LOG v1.0 | ||
| For Annex III point 1(a) systems, each use has a logged start date and time and end date and time, and the two timestamps bound the observed session (Art. 12(3)(a)) | PR-12-LOG v1.0 | ||
| For Annex III point 1(a) systems, each match records the identifier of the reference database checked and the input data that produced the match (Art. 12(3)(b) and (c)) | PR-12-LOG v1.0 | ||
| For Annex III point 1(a) systems, each verified result records the identification of the natural persons who performed the verification under Art. 14(5) (Art. 12(3)(d)) | PR-12-LOG v1.0 | ||
| Art. 13 | The accuracy level and metrics declared in the instructions for use under Art. 13(3)(b)(ii) match the result of a test on a held-out set under the declared conditions of use. | PR-11-DOC v1.0 |
|
| The robustness and cybersecurity levels declared in the instructions for use under Art. 13(3)(b)(ii) match the result of perturbation and adversarial tests under the declared conditions of use. | PR-11-DOC v1.0 | ||
| The known and foreseeable circumstances declared under Art. 13(3)(b)(ii) and (iii) reproduce the declared change in performance or the declared risk when the system runs under those circumstances. | PR-11-DOC v1.0 | ||
| The performance declared for specific persons or groups under Art. 13(3)(b)(v) matches the disaggregated results recorded in the technical documentation. | PR-11-DOC v1.0 | ||
| Each statement in the instructions for use on intended purpose, input data specifications, pre-determined changes, human oversight measures and log mechanisms (Art. 13(3)(b)(i), (b)(vi), (c), (d) and (f)) matches the corresponding section of the Annex IV technical documentation. | PR-11-DOC v1.0 | ||
| Art. 14 | Human-machine interface tools are present, and a natural person assigned to oversight can operate them while the system is in use (Art. 14(1)). | PR-14-OVR v1.0 |
|
| The assigned person can monitor operation from the interface: an anomaly, dysfunction or unexpected performance injected during the test is visible to the assigned person and can be addressed (Art. 14(4)(a)). | PR-14-OVR v1.0 | ||
| A measure against automation bias is present in the interface or the workflow and has an observable effect, for example the output is not pre-filled as the decision and the assigned person records an active choice (Art. 14(4)(b)). | PR-14-OVR v1.0 | ||
| Interpretation tools or methods are available to the assigned person and return a result for a sampled output (Art. 14(4)(c)). | PR-14-OVR v1.0 | ||
| The assigned person can decide not to use the system for a particular case, and can disregard, override or reverse the output; the recorded decision then carries the person's choice and not the system output (Art. 14(4)(d)). | PR-14-OVR v1.0 | ||
| A stop button or a similar procedure exists; when it is used the system comes to a halt, and the state after the stop matches the safe state the provider declares (Art. 14(4)(e)). | PR-14-OVR v1.0 | ||
| For an Annex III point 1(a) biometric identification system, the workflow blocks any action or decision until two natural persons have separately verified and confirmed the identification; one confirmation, or two confirmations from the same person, does not release the action (Art. 14(5)). | PR-14-OVR v1.0 | ||
| Art. 15 | Accuracy on a held-out test set, measured with the metrics declared in the instructions for use, is at or above the declared level (Art. 15(1) and 15(3)) | PR-15-ROB v1.0 |
|
| Accuracy holds on a distribution-shifted set and across the released model versions, so that the system performs consistently throughout its lifecycle (Art. 15(1)) | PR-15-ROB v1.0 | ||
| Outputs stay within the declared performance under input perturbation, noise, malformed input, edge cases and inconsistent signals from natural persons or connected systems (Art. 15(4), first subparagraph) | PR-15-ROB v1.0 | ||
| The technical redundancy, backup or fail-safe solution takes over and holds the declared behaviour when a component fault is injected, where such a solution exists (Art. 15(4), second subparagraph) | PR-15-ROB v1.0 | ||
| For a system that continues to learn, feeding its own possibly biased outputs back as inputs does not move performance beyond the declared level, and the mitigation measure activates as documented (Art. 15(4), third subparagraph) | PR-15-ROB v1.0 | ||
| The system resists each attack class named in Art. 15(5), third subparagraph: data poisoning of the training set, model poisoning of pre-trained components, adversarial examples or model evasion, confidentiality attacks such as model extraction and membership inference, and exploitation of model flaws, with prompt injection included for systems built on a language model | PR-15-ROB v1.0 | ||
| The cybersecurity measures detect, respond to, resolve and control a simulated unauthorised attempt to alter the use, outputs or performance of the system (Art. 15(5), first and third subparagraphs) | PR-15-ROB v1.0 | ||
| Art. 17 | Each released version of the high-risk AI system has a modification record that states whether the change is a substantial modification and which conformity assessment step follows (Art. 17(1)(a)); verified by record sampling | PR-17-QMS v1.0 |
|
| Design verification records exist for the current version and trace each verified item to a design input (Art. 17(1)(b)); verified by record sampling | PR-17-QMS v1.0 | ||
| Quality control records for the current release show the checks and outcomes that the development procedure names (Art. 17(1)(c)); verified by record sampling | PR-17-QMS v1.0 | ||
| Examination, test and validation records exist for the current version at the frequency the procedure states, and each record names the version tested (Art. 17(1)(d)); verified by record sampling against the stated frequency | PR-17-QMS v1.0 | ||
| Dataset records for the training, validation and testing sets show the acquisition, labelling, storage and retention steps that the data management procedure names (Art. 17(1)(f)); verified by record sampling | PR-17-QMS v1.0 | ||
| A simulated serious incident walked through the reporting procedure reaches the reporting step with the market surveillance authority and the Art. 73 timeframe identified (Art. 17(1)(i)); verified by procedure walk-through | PR-17-QMS v1.0 | ||
| A sample of the records named in the record-keeping procedure is retrievable for the current version and for one earlier version within the retention period the procedure states (Art. 17(1)(k)); verified by retrieval test | PR-17-QMS v1.0 | ||
| Each responsibility in the accountability framework is held by a named role, and the role holder confirms the responsibility in a walk-through (Art. 17(1)(m)); verified by role walk-through | PR-17-QMS v1.0 |
Art. 43, Annex VI and Annex VII set no numeric threshold; each annex point has a binary criterion drawn from its text, and the protocol's own severity scale, agreed at the start of the engagement, has no source in the Act or in a standard. For the sampled re-tests the pass criterion comes from the protocol that tests the record, following a six-level hierarchy: a harmonised standard cited in the Official Journal, of which none exists as of the anchor date; a JTC 21 draft standard named with its stage; an ISO/IEC standard or technical report; a Commission guideline or code of practice; NIST AI RMF or a sector framework; and the provider's justified state of the art. Where no published threshold exists, the provider sets the target and justifies it against the context of use, as Art. 9(8) requires for prior defined metrics and probabilistic thresholds, and the threshold-source step of PR-43-MOCK v1.0 records the declared threshold, its source and whether the result meets it. A sampled test with no declared threshold is recorded as a blocking finding against Art. 9(8), because the result cannot be judged. No figure on this page is attributed to the Act.
Protocols used
One protocol runs the service; its version and derivation are on its page.
- PR-43-MOCK v1.0Pre-conformity mock assessment protocolv1.0
PR-43-MOCK v1.0 covers the route rules in Art. 43(1) to 43(4), each point of Annex VI and Annex VII, the sampled re-run of testable elements across the eight records, and the presence checks for the declaration of conformity, the CE marking plan and the registration record. It produces a route selection record, a dry-run review report, a findings register, an evidence-pack index, a remediation plan, the two presence checklists and a modification register. It does not show that every element passes, since the sample says nothing about the elements outside it; it observes the quality management system at one point in time, while Annex VII point 5.3 provides for periodic audits; and it does not issue or sign the declaration, affix the marking or submit the registration, which remain with the provider under Art. 47, Art. 48 and Art. 49.
Engagement shape
The service is fixed-scope, scoped per request, and follows the obligation register from the Classification and readiness assessment where one exists: the classification memo names the Annex III point or the Annex I Section A entry that selects the route. The client provides the technical documentation under Art. 11 with Annex IV, the quality management system documentation under Art. 17(1), the test reports and logs the documentation names, the instructions for use, the draft declaration of conformity, the CE marking plan, the registration record or draft, the change and release records, and read access to the data sets and to the system for the sampled re-tests. The work follows the order of the protocol. Preparation records the system, the classification basis, the standards applied and the route, agrees the layouts and the severity scale, logs access and freezes the version. Execution works through the quality management system, the technical documentation, the sample, the data-set and model access, the declared-versus-measured comparison, the design and post-market monitoring consistency, the three presence checks, the surveillance provisions and the modification register, point by point. Analysis gives each finding a type, a severity and a reference, fills the status column of the index, records the threshold source of each sampled test and writes the dry-run conclusion. Reporting issues the register, the index, the plan and the review report with the scope statement that no presumption of conformity follows from a dry run, since no harmonised standard is cited in the Official Journal as of the anchor date.
What we cannot do
- We do not perform conformity assessments where a notified body is required.
- We do not issue CE marking or the EU Declaration of Conformity.
- We are not a notified body.
- We do not give legal opinions on classification or liability.
- We test against requirements, prepare evidence, and structure documentation; compliance decisions remain with the provider or deployer, their notified body, and their counsel.
For this service the dry run is a rehearsal of the Annex VI or Annex VII route; no certificate, no declaration and no CE marking follow from it, and a notified body or a market surveillance authority can reach a different result on the same evidence. The remediation plan orders the findings by severity and by the annex point each blocks; the decision to enter the real assessment rests with the provider.
What the client must have ready
The documentation and process elements below are what the annex points are checked against; the evidence-pack index records each as present, partial or absent.
Art. 9Risk management system
- The risk management system is documented in a risk management file that records the risks identified, the estimation and evaluation method, the measures adopted and the residual risk judgement (Art. 9(1) and 9(5)). documentation
- Each hazard is recorded with the intended purpose and the reasonably foreseeable misuse conditions under which it was identified and evaluated (Art. 9(2)(a) and 9(2)(b)). documentation
- For each risk, the file states whether design, development or adequate technical information can reasonably mitigate or eliminate it, which fixes the scope of Art. 9 for that risk (Art. 9(3)). documentation
- The interaction analysis records how the measures taken under Art. 10 to Art. 15 affect each other and the balance the provider chose (Art. 9(4)). documentation
- The acceptance criterion for the residual risk of each hazard and for the overall residual risk is stated before testing, with its rationale (Art. 9(5)). documentation
- The Art. 13 information and any deployer training are described together with the technical knowledge, experience, education and training expected of the deployer and the presumable context of use (Art. 9(5)(c)). documentation
- Where testing in real-world conditions takes place, the plan and the conditions under Art. 60 are documented (Art. 9(7)). documentation
- The prior defined metrics and probabilistic thresholds are recorded before the test runs, with the reason each is appropriate to the intended purpose (Art. 9(8)). documentation
- Test records identify the system version tested and show that a full test run took place before placing on the market or putting into service (Art. 9(8)). documentation
- The consideration of adverse impact on persons under the age of 18 and other vulnerable groups is recorded with its outcome (Art. 9(9)). documentation
- Where the provider combines Art. 9 with a risk management procedure under other Union law, a mapping shows where each aspect of Art. 9(1) to 9(9) is covered (Art. 9(10)). documentation
- The risk management system is established, implemented and maintained as a continuous iterative process that is planned and run through the entire lifecycle of the system (Art. 9(1) and 9(2)). process
- The risk management system is reviewed and updated regularly and systematically, with a defined trigger and a record of each review (Art. 9(2)). process
- Data from the post-market monitoring system under Art. 72 is analysed and other risks that arise are evaluated in the risk management cycle (Art. 9(2)(c)). process
- Risk management measures are adopted for the identified risks in the order of Art. 9(5): elimination or reduction by design and development, then mitigation and control measures, then information and training for deployers (Art. 9(2)(d) and 9(5)). process
- Testing takes place at any time during the development process as appropriate, and is repeated before placing on the market or putting into service (Art. 9(8)). process
- A provider subject to internal risk management requirements under other Union law may run the Art. 9 aspects inside that procedure (Art. 9(10)). process
Art. 10Data and data governance
- Statement of whether the system uses techniques involving the training of AI models, which determines whether the criteria apply to all data sets or to the testing data sets only (Art. 10(1) and 10(6)) documentation
- Design choices record for the data sets (Art. 10(2)(a)) documentation
- Data collection and origin record, including the original purpose of collection for personal data (Art. 10(2)(b)) documentation
- Data-preparation record listing the annotation, labelling, cleaning, updating, enrichment and aggregation operations (Art. 10(2)(c)) documentation
- Statement of assumptions about the information the data are supposed to measure and represent (Art. 10(2)(d)) documentation
- Assessment of the availability, quantity and suitability of the data sets that are needed (Art. 10(2)(e)) documentation
- Record of identified data gaps or shortcomings and how they are addressed (Art. 10(2)(h)) documentation
- Justification of the representativeness, error and completeness targets against the intended purpose, and a statement of whether the characteristics are met per data set or by a combination of data sets (Art. 10(3)) documentation
- Record of the reasons why processing special categories of personal data was strictly necessary and why other data, including synthetic or anonymised data, could not achieve the objective, in the records of processing activities (Art. 4a(1)(a) and (f)) documentation
- Access control documentation for special-category data, listing the authorised persons and their confidentiality obligations (Art. 4a(1)(c)) documentation
- Retention period statement for special-category data processed for bias detection (Art. 4a(1)(e)) documentation
- Data governance and management practices operated across the lifecycle for the training, validation and testing data sets (Art. 10(2)) process
- Bias examination repeated when a data set changes and, where data outputs influence inputs for future operations, at each operation cycle (Art. 10(2)(f)) process
- Bias detection, prevention and mitigation measures applied and maintained as identified biases change (Art. 10(2)(g)) process
- Data gap and shortcoming identification repeated at each data set update (Art. 10(2)(h)) process
- Deletion of special-category data at the earlier of bias correction and the end of the retention period (Art. 4a(1)(e)) process
- Where a deployer or a provider of another AI system processes special-category data under Art. 4a(2), application of all the Art. 4a(1) conditions and safeguards to that processing (Art. 4a(2)) process
Art. 11Technical documentation
- A technical documentation set exists that contains, at a minimum, the elements set out in Annex IV, in a clear and comprehensive form for national competent authorities and notified bodies (Art. 11(1)) documentation
- For an SME, including a start-up, or an SMC that opts for the simplified manner, the documentation uses the Commission's simplified technical documentation form (Art. 11(1) third subparagraph) documentation
- For a system related to an Annex I Section A product, a single documentation set holds the Art. 11(1) information and the information required under the product legislation (Art. 11(2)) documentation
- General description: intended purpose with context and conditions of use, provider name, system version and relation to previous versions, interaction with external hardware, software and other AI systems, software and firmware versions and update requirements, forms of placing on the market, target hardware, product photographs or illustrations where the system is a product component, and a basic description of the deployer user interface (Annex IV section 1(a) to 1(g)) documentation
- Instructions for use for the deployer (Annex IV section 1(h)) documentation
- Development methods and steps, including third-party pre-trained systems or tools and how they were used, integrated or modified (Annex IV section 2(a)) documentation
- Design specifications: general logic, algorithms, key design choices with rationale and assumptions, classification choices, optimisation target and parameter relevance, expected output and output quality, and trade-off decisions taken to meet Chapter III Section 2 (Annex IV section 2(b)) documentation
- Architecture description and the computational resources used to develop, train, test and validate the system (Annex IV section 2(c)) documentation
- Datasheets for the training data sets with training methodologies and techniques, provenance, scope, main characteristics, acquisition and selection, labelling and cleaning (Annex IV section 2(d)) documentation
- Assessment of the human oversight measures under Art. 14 and of the technical measures for output interpretation under Art. 13(3)(d) (Annex IV section 2(e)) documentation
- Detailed description of pre-determined changes and the technical solutions that keep the system within Chapter III Section 2 (Annex IV section 2(f)) documentation
- Description of the validation and testing procedures, the validation and testing data with their main characteristics, and the metrics used for accuracy, robustness, other Chapter III Section 2 requirements and potentially discriminatory impacts (Annex IV section 2(g)) documentation
- Description of the cybersecurity measures put in place (Annex IV section 2(h)) documentation
- Capabilities and limitations in performance, foreseeable unintended outcomes, sources of risk to health and safety, fundamental rights and discrimination, human oversight measures with output interpretation measures, and input data specifications (Annex IV section 3) documentation
- Description of the appropriateness of the performance metrics for the specific system (Annex IV section 4) documentation
- Detailed description of the risk management system under Art. 9 (Annex IV section 5) documentation
- List of the applied harmonised standards cited in the Official Journal, or, where none is applied, a detailed description of the solutions adopted and a list of the other standards and technical specifications applied (Annex IV section 7) documentation
- Copy of the EU declaration of conformity under Art. 47 (Annex IV section 8) documentation
- Detailed description of the post-market monitoring system under Art. 72, including the post-market monitoring plan under Art. 72(3) (Annex IV section 9) documentation
- The technical documentation is drawn up before the system is placed on the market or put into service (Art. 11(1)) process
- The technical documentation is kept up to date over the lifecycle of the system (Art. 11(1)) process
- Test logs and test reports are dated and signed by the responsible persons when they are produced, including for each pre-determined change (Annex IV section 2(g)) process
- Relevant changes made to the system through its lifecycle are recorded as they occur (Annex IV section 6) process
Art. 12Record-keeping
- Log schema and event catalogue: the event set, the fields per event, the format and the storage location, in the technical documentation under Annex IV section 3 (Art. 12(1)) documentation
- Traceability rationale: a statement of the level of traceability appropriate to the intended purpose and how the chosen event set achieves it (Art. 12(2)) documentation
- Mapping from each risk indicator in the risk management file and each substantial-modification trigger to the log event that captures it (Art. 12(2)(a)) documentation
- Mapping from each data point in the post-market monitoring plan under Art. 72 to the log event that supplies it (Art. 12(2)(b)) documentation
- Description in the instructions for use of the mechanisms that allow the deployer to collect, store and interpret the logs (Art. 12(2)(c) and Art. 13(3)(f)) documentation
- For Annex III point 1(a) systems, the field specification for the period of use, the reference database identifier, the matched input data and the verifier identification (Art. 12(3)) documentation
- Retention of the logs under the provider's control for the period set in Art. 19, and of the logs under the deployer's control for the period set in Art. 26(6) process
- Change control that re-verifies the logging capability after each substantial modification or version change over the lifetime of the system (Art. 12(1)) process
- Review of the event set whenever the intended purpose, the risk management file under Art. 9 or the post-market monitoring plan under Art. 72 changes (Art. 12(2)) process
- Periodic review of the logs as an input to the post-market monitoring system under Art. 72 (Art. 12(2)(b)) process
Art. 13Transparency and provision of information to deployers
- Instructions for use exist in an appropriate digital format or otherwise and accompany the system (Art. 13(2)). documentation
- The information in the instructions for use is concise, complete, clear, relevant, accessible and comprehensible to deployers (Art. 13(2)). documentation
- The instructions for use state the identity and contact details of the provider and, where applicable, of its authorised representative (Art. 13(3)(a)). documentation
- The instructions for use state the intended purpose (Art. 13(3)(b)(i)). documentation
- The instructions for use declare the level of accuracy with its metrics, robustness and cybersecurity against which the system was tested and validated, and the known and foreseeable circumstances that affect that level (Art. 13(3)(b)(ii)). documentation
- The instructions for use list the known or foreseeable circumstances under the intended purpose or reasonably foreseeable misuse that may lead to risks to health, safety or fundamental rights (Art. 13(3)(b)(iii)). documentation
- Where applicable, the instructions for use describe the technical capabilities and characteristics that provide information relevant to explain the output (Art. 13(3)(b)(iv)). documentation
- When appropriate, the instructions for use state the performance regarding the specific persons or groups on which the system is intended to be used (Art. 13(3)(b)(v)). documentation
- When appropriate, the instructions for use give specifications for the input data and relevant information on the training, validation and testing data sets (Art. 13(3)(b)(vi)). documentation
- Where applicable, the instructions for use give information that enables deployers to interpret the output and use it appropriately (Art. 13(3)(b)(vii)). documentation
- The instructions for use state the changes to the system and its performance that the provider pre-determined at the initial conformity assessment, if any (Art. 13(3)(c)). documentation
- The instructions for use describe the human oversight measures under Art. 14 and the technical measures that facilitate interpretation of the outputs (Art. 13(3)(d)). documentation
- The instructions for use state the computational and hardware resources needed, the expected lifetime, and the maintenance and care measures with their frequency, including software updates (Art. 13(3)(e)). documentation
- Where relevant, the instructions for use describe the mechanisms that allow deployers to collect, store and interpret the logs under Art. 12 (Art. 13(3)(f)). documentation
- Transparency of operation is a design and development objective carried through the lifecycle of the system, at a type and degree appropriate to the obligations of the provider and the deployer (Art. 13(1)). process
- The maintenance and care measures declared under Art. 13(3)(e), including their frequency and software updates, are carried out over the expected lifetime of the system. process
- The instructions for use are revised when a pre-determined change under Art. 13(3)(c) is applied or when a re-test changes the declared level under Art. 13(3)(b)(ii). process
Art. 14Human oversight
- The measures built into the system are identified before it is placed on the market or put into service, and are described in the technical documentation (Art. 14(3)(a); Annex IV section 2(e)). documentation
- The measures appropriate for the deployer to implement are identified before placing on the market or putting into service, and are stated in the instructions for use (Art. 14(3)(b); Art. 13(3)(d)). documentation
- A written rationale links each oversight measure to the risks, the level of autonomy and the context of use of the system (Art. 14(3)). documentation
- The risks to health, safety or fundamental rights under the intended purpose and under reasonably foreseeable misuse that persist despite the other Section 2 requirements are listed, with the oversight measure that addresses each (Art. 14(2), read with the Art. 9 risk management file). documentation
- The relevant capacities and limitations of the system are stated so that the assigned person can understand them (Art. 14(4)(a); Art. 13(3)(b)). documentation
- The automation bias risk and the measures against it are described for the deployer, in particular where the system provides information or recommendations for decisions taken by natural persons (Art. 14(4)(b)). documentation
- The interpretation tools and methods available to the assigned person are described (Art. 14(4)(c)). documentation
- Where a law enforcement, migration, border control or asylum system does not apply the two-person verification, the Union or national law that treats the requirement as disproportionate is identified (Art. 14(5) second subparagraph). documentation
- The oversight measures stay effective during the whole period in which the system is in use; a change to the system or to its interface is followed by a repeat of the oversight verification (Art. 14(1)). process
- The deployer implements the measures the provider identified for the deployer and assigns oversight to natural persons (Art. 14(3)(b), read with Art. 26(1) and (2)). process
- Persons assigned to oversight are kept aware of automation bias during use; the awareness measure is repeated over time rather than given once (Art. 14(4)(b)). process
- For an Annex III point 1(a) biometric identification system, the two natural persons who verify an identification have the necessary competence, training and authority, and the assignment is maintained for each person who takes the role (Art. 14(5)). process
Art. 15Accuracy, robustness and cybersecurity
- Declaration of the accuracy levels and the relevant accuracy metrics in the instructions for use, with the metric definitions, the test-set description and the measurement conditions (Art. 15(3)) documentation
- Rationale for the level of accuracy, robustness and cybersecurity chosen as appropriate for the intended purpose, with the source of each target (Art. 15(1)) documentation
- Description of the technical and organisational measures taken for resilience to errors, faults and inconsistencies (Art. 15(4), first subparagraph) documentation
- Description of the technical redundancy solution and the backup or fail-safe plan where one is used (Art. 15(4), second subparagraph) documentation
- Design description of how a continuously learning system limits feedback loops and the mitigation measures that address them (Art. 15(4), third subparagraph) documentation
- Cybersecurity risk assessment that shows the technical solutions are appropriate to the relevant circumstances and the risks, and that records, for each attack class in Art. 15(5), whether a measure applies and the reason where one is omitted as not appropriate (Art. 15(5), second and third subparagraphs) documentation
- Re-measurement of accuracy, robustness and cybersecurity after each model, data or operating-environment change, with a defined trigger, so that consistent performance is shown across the lifecycle (Art. 15(1)) process
- Operation of the organisational measures for resilience: fault handling, incident escalation and the operating procedures that respond to errors and inconsistencies in service (Art. 15(4), first subparagraph) process
- Operation of the feedback-loop mitigation measures while the system continues to learn, with periodic review of learned behaviour (Art. 15(4), third subparagraph) process
- Ongoing prevention, detection, response, resolution and control of AI-specific attacks: vulnerability management, threat monitoring and the response path for a detected attack (Art. 15(5)) process
Art. 17Quality management system
- The quality management system is documented in written policies, procedures and instructions (Art. 17(1)) documentation
- A strategy for regulatory compliance names the conformity assessment procedure the provider follows and the procedure for the management of modifications (Art. 17(1)(a)) documentation
- Techniques, procedures and systematic actions for design, design control and design verification are written (Art. 17(1)(b)) documentation
- Techniques, procedures and systematic actions for development, quality control and quality assurance are written (Art. 17(1)(c)) documentation
- Examination, test and validation procedures state what is carried out before, during and after development, and the frequency of each (Art. 17(1)(d)) documentation
- A list of the technical specifications and standards applied, and, where harmonised standards are not applied in full or do not cover a Chapter III Section 2 requirement, a statement of the means used for that requirement (Art. 17(1)(e)) documentation
- Data management procedures cover data acquisition, collection, analysis, labelling, storage, filtration, mining, aggregation and retention (Art. 17(1)(f)) documentation
- The post-market monitoring plan that Art. 72(3) requires is part of the quality management system documentation (Art. 17(1)(h)) documentation
- Serious incident reporting procedures name the reporting path, the authority and the Art. 73 timeframes (Art. 17(1)(i)) documentation
- Procedures for the handling of communication with authorities, notified bodies, other operators, customers and other interested parties are written (Art. 17(1)(j)) documentation
- Record-keeping procedures name the records kept, where they are kept and the retention period of each (Art. 17(1)(k)) documentation
- Resource management, including security-of-supply related measures, is described (Art. 17(1)(l)) documentation
- The accountability framework sets out the responsibilities of the management and other staff for each Art. 17(1) aspect (Art. 17(1)(m)) documentation
- A proportionality rationale records how the size of the organisation, including SME, start-up or SMC status, shaped the implementation and shows that the degree of rigour is kept (Art. 17(2)) documentation
- Where a sectoral quality management system is used, a mapping from each Art. 17(1) aspect to the element of the sectoral system that holds it (Art. 17(3)) documentation
- Where Art. 17(4) applies, a mapping from the Union financial services governance rules to the quality management system obligation, with points (g), (h) and (i) documented separately (Art. 17(4)) documentation
- Where an SME uses the simplified manner, a record of the SME status without partner or linked enterprises and of the elements simplified under the Commission guidelines (Art. 63(1)) documentation
- The quality management system operates across the lifecycle and is kept current after each modification of the system (Art. 17(1)) process
- Modifications to the high-risk AI system are managed under the written procedure before release (Art. 17(1)(a)) process
- Quality assurance activities run during development as the procedure states (Art. 17(1)(c)) process
- Examination, test and validation are repeated after development at the stated frequency (Art. 17(1)(d)) process
- The risk management system under Art. 9 runs as part of the quality management system (Art. 17(1)(g)) process
- The post-market monitoring system is set up, implemented and maintained in accordance with Art. 72 (Art. 17(1)(h)) process
- Serious incidents are reported to the market surveillance authority in accordance with Art. 73 (Art. 17(1)(i)) process
- Communication with authorities, notified bodies, other operators and customers follows the written procedure (Art. 17(1)(j)) process
- Records are kept under the record-keeping procedure for the period set in Art. 18 (Art. 17(1)(k)) process
- Resources, including security of supply, are managed under the written measures (Art. 17(1)(l)) process
- Responsibilities are assigned to roles and kept current as management and staff change (Art. 17(1)(m)) process
- Risk management, post-market monitoring and serious incident reporting continue for a financial institution that relies on Art. 17(4) (Art. 17(4)) process
Inputs take the formats listed on the PR-43-MOCK page: each document with a version identifier, test reports dated and signed, the log export in the schema the documentation names, the declaration in the Annex V structure, the registration record as a structured export of the Annex VIII Section A fields, the change log with version identifiers and change classification, and system and data-set access as read-only API or environment access with the safeguards the provider names.
Questions
What is the difference between Annex VI and Annex VII conformity assessment?
Annex VI is internal control: the provider itself checks its quality management system against Art. 17, that the technical documentation shows each Chapter III Section 2 requirement is met, and that the design process matches the documentation. No notified body takes part in Annex VI. Annex VII is assessment of the quality management system and the technical documentation by a notified body, with data-set access, further tests, surveillance audits and a certificate on a positive decision. PR-43-MOCK steps 3 to 6 record the route under Art. 43(1) and 43(2).
When do you need a notified body under the EU AI Act?
Under Art. 43(1), an Annex III point 1 provider may choose Annex VI only where it applied harmonised standards or common specifications in full; otherwise Annex VII with a notified body applies. No harmonised standard is cited in the Official Journal as of the anchor date, so the Art. 43(1) condition is not met. Under Art. 43(2), Annex III points 2 to 8 systems use Annex VI without a notified body. Under Art. 43(3), an Annex I Section A product follows its sectoral procedure and notified body.
Which AI systems require conformity assessment under the EU AI Act?
Art. 43 applies to every high-risk AI system: an Annex III system classified under Art. 6(2), unless the provider concludes under Art. 6(3) that the system is not high-risk and documents the conclusion under Art. 6(4), and an Annex I Section A product classified under Art. 6(1). Art. 16(f) requires the provider to have the system undergo the procedure before placing it on the market, and Art. 43(4) starts a new assessment after a substantial modification. A system with only Art. 50 duties has no conformity assessment duty.
How does a provider prepare for a notified body assessment under the AI Act?
Annex VII point 3.1 lists what the notified body receives with the quality management system application: the provider identity, the systems covered, the technical documentation of each, the Art. 17 documentation, the procedure that keeps the system adequate, and a declaration that no other notified body received the application. Annex VII points 4.3 to 4.5 give the notified body access to the data sets, further tests and, as a last resort, the model. PR-43-MOCK steps 12 to 20 check each item and record every absence in the findings register.
What does an EU AI Act internal audit cover?
The Act names no internal audit. Annex VI point 1 requires the provider to run the internal control of points 2 to 4 itself, and Art. 17(1)(a) requires the quality management system to cover the conformity assessment procedures. An internal audit built on Annex VI points 2 to 4 is a dry run: the quality management system against Art. 17, the technical documentation against Chapter III Section 2, the design process against the documentation. PR-43-MOCK runs the Annex VI dry run; existing audit records enter steps 11 to 14.
Every inquiry receives a reply within 24 hours. Scope and availability are discussed per request.