{"answer_id":"praxikon:eu:ai-act:answer:testomgeving-en-praktijktest","canonical_page":"https://www.praxikon.com/en/antwoord/testomgeving-en-praktijktest","query":"Can we test our AI system before it goes to market?","lang":"en","view":"full","mode":"scenario","question":"Can we test our AI system before it goes to market?","situation":"You want to develop, train and validate an AI system before placing it on the market, in a supervised sandbox or under real world conditions outside the laboratory.","likely_role":"Provider (you are preparing market placement)","note":"There are two routes and they are not the same. The AI regulatory sandbox of Article 57 is a controlled environment supervised by the competent authority, and every member state must provide at least one. Testing in real world conditions under Article 60 happens outside that environment, with an approved testing plan, informed consent from participants and a limit in time. Both are voluntary routes: they replace no obligation, but they do give you supervision and a written record of what you tested.","matched_terms":[],"dataset":{"id":"praxikon:sys:registry:dataset:ai-act-implementation-graph","version":"2.1.0","schema_version":"1.4.0","effective_at":"2026-08-08T00:00:00.000Z","known_at":"2026-08-14T00:00:00.000Z","last_reviewed_at":"2026-08-08T00:00:00.000Z","licence":"https://www.praxikon.com/nl/legal/terms","canonical_url":"https://www.praxikon.com/api/v1/entities"},"obligations":[{"slug":"article-60-real-world-testing","label":"Article 60: testing in real world conditions outside a sandbox","summary":"If you want to test an Annex III high-risk AI system with real people and real outcomes before placing it on the market, a full regime applies: a plan, prior approval by the market surveillance authority, registration, informed consent and a maximum duration.","legal_status":"applicable","deadline_at":"2026-08-02T00:00:00.000Z","human_page":"https://www.praxikon.com/en/verplichtingen/article-60-real-world-testing","api":"https://www.praxikon.com/api/v1/obligations?lang=en","official_source":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","citations":[{"kind":"official_fact","statement":"Article 60(1) provides that testing of high-risk AI systems in real world conditions outside AI regulatory sandboxes may be conducted by providers or prospective providers of high-risk AI systems listed in Annex III, in accordance with that Article and the real-world testing plan, without prejudice to the prohibitions under Article 5. The Commission specifies the detailed elements of that plan by implementing act. The third subparagraph of paragraph 1 provides that the paragraph is without prejudice to Union or national law on the testing in real world conditions of high-risk AI systems related to products covered by the Union harmonisation legislation listed in Annex I. Article 60(2) allows providers or prospective providers to test at any time before placing on the market or putting into service, on their own or in partnership with one or more deployers or prospective deployers. Article 60(3) provides that such testing is without prejudice to any ethical review required by Union or national law. Article 60(4), point (f), caps the duration: no longer than necessary to achieve its objectives and in any case no longer than six months, which may be extended by an additional six months subject to prior notification to the market surveillance authority with an explanation of the need. Article 60(4), point (g), requires that subjects belonging to vulnerable groups due to age or disability are appropriately protected. Article 60(9) expressly states that the provider or prospective provider remains fully subject to applicable Union and national law on any damage caused in the course of their testing in real world conditions. Chapter VI, which contains Article 60, is not among the exceptions in Article 113 and applies since 2 August 2026.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 60(1)-(4), Article 60(9), Article 113","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"editorial_interpretation","statement":"Many organisations call what they do a pilot and assume that keeps them outside the Regulation. Article 60 shows that this does not hold once you test an Annex III system in real world conditions with real people and real outcomes. A full regime then applies: a plan, prior approval, registration with a Union-wide unique single identification number, informed consent, and a hard six-month clock with a maximum six-month extension. The heaviest requirement in practice is Article 60(4), point (k): the predictions, recommendations or decisions of the system must be capable of being effectively reversed and disregarded. If you are testing a selection, scoring or triage system whose output feeds straight into the workflow with nobody able to reverse it, your design does not qualify, however careful your consent form is. Note the timing too, because it is commercially interesting. Chapter VI applies since 2 August 2026, while the core obligations for standalone Annex III systems only apply from 2 December 2027. The testing route is therefore open before the requirements themselves bite, and that is exactly the window in which to validate your design rather than rebuild it later. Finally, Article 60(3) leaves any ethical review required under other law fully in place, and Article 60(9) expressly states that you remain fully subject to the applicable law on damage caused during the testing.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 60(1)-(4), Article 60(9), Article 113","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"recommended_action","statement":"Inventory which running or planned trials are in fact real-world testing: real users, real data, outputs that feed into the workflow. Test those first against Article 60(4), point (k): can the output genuinely be reversed and disregarded? If not, redesign the trial before you submit anything. Then choose deliberately between two routes: supervised testing inside a sandbox under Article 57(5) and Article 58(4), or outside a sandbox under Article 60. Plan the six months realistically and decide in advance at which point you will request an extension, since that requires prior notification with a reasoned explanation. Check whether an ethical review is mandatory in your domain and start it in parallel, because Article 60(3) does not exempt you from it.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 60(1)-(4), Article 60(9), Article 113","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"}]},{"slug":"article-57-regulatory-sandboxes","label":"Article 57: AI regulatory sandboxes","summary":"Member States must provide at least one national AI regulatory sandbox. For you this is a voluntary route: you develop, train, test and validate an innovative AI system in a controlled, supervised environment under a plan agreed with the competent authority, before placing it on the market or putting it into service.","legal_status":"applicable","deadline_at":"2026-08-02T00:00:00.000Z","human_page":"https://www.praxikon.com/en/verplichtingen/article-57-regulatory-sandboxes","api":"https://www.praxikon.com/api/v1/obligations?lang=en","official_source":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","citations":[{"kind":"official_fact","statement":"Article 57(1) requires Member States to ensure that their competent authorities establish at least one AI regulatory sandbox at national level. The text of Regulation (EU) 2024/1689 as published in the Official Journal provides that the sandbox shall be operational by 2 August 2026, and that wording still stands unchanged on 9 August 2026 in the article text published by the European Commission. The timeline the European Commission publishes after the Digital Omnibus, Regulation (EU) 2026/1744, entered into force on 27 July 2026 places the milestone at 2 August 2027 instead, stating \"Member States should have at least one AI regulatory sandbox per country operational\". The consolidated text of Article 57(1) after the Omnibus has not been verified at article level; rely on the consolidated version on EUR-Lex for the exact date. Chapter VI, which contains Article 57, is not among the exceptions in Article 113 and therefore applies since 2 August 2026. Article 57(1) further provides that the obligation may also be fulfilled by participating in an existing sandbox, in so far as that participation provides an equivalent level of national coverage. Article 57(5) defines the sandbox as a controlled environment that fosters innovation and facilitates the development, training, testing and validation of innovative AI systems for a limited time, pursuant to a specific sandbox plan agreed between the provider or prospective provider and the competent authority, before the system is placed on the market or put into service. Such sandboxes may include testing in real world conditions supervised therein. Article 57(15) requires the AI Office to make publicly available, and keep up to date, a list of planned and existing sandboxes.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 57(1)-(17), Article 58, Article 113; as amended by Regulation (EU) 2026/1744","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"editorial_interpretation","statement":"The common mistake is to read the sandbox as a delay or an exemption. It is neither. Article 57(11) leaves the competent authority's supervisory and corrective powers fully intact and expressly empowers it to suspend your testing or your participation, temporarily or permanently, where risks cannot be effectively mitigated. A sandbox is useful for exactly one profile: you are building an AI system that is likely to fall under Annex III, you are uncertain about its classification or about how to meet Chapter III, Section 2, and you want that uncertainty resolved before you go to market. If you are purely a deployer buying a system, the sandbox is not your route: at most you can join as a partner of the provider under Article 58(2), point (b). Second trap: 2 August 2027 is a duty on the Member State, not on you. The date on this item is therefore 2 August 2026: from that moment Chapter VI applies and the route is open to you. You cannot hang your own preparation on that date, and the date says nothing about whether your national sandbox will actually have capacity by then. Third trap: people confuse the Article 57 sandbox with Article 60 real-world testing. These are two distinct routes with different conditions. Real-world testing can take place inside the sandbox (Article 57(5) and Article 58(4)) or outside it under Article 60, and the safeguards differ.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 57(1)-(17), Article 58, Article 113; as amended by Regulation (EU) 2026/1744","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"recommended_action","statement":"First establish your role: only providers and prospective providers can enter on their own. Then use the AI Office's published list of planned and existing sandboxes (Article 57(15)) to find the sandbox open to you, in your Member State or jointly with others. Before you apply, state exactly which uncertainty you want resolved, ideally focused on classification under Article 6 or on a specific requirement in Chapter III, Section 2. Plan for the three-month decision period that Article 58(2), point (a), requires the implementing acts to ensure, build it into your product timeline, and ask the competent authority in writing which period it currently applies. On entry, record in writing what the sandbox plan covers, and on exit request both the written proof and the exit report.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 57(1)-(17), Article 58, Article 113; as amended by Regulation (EU) 2026/1744","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"}]},{"slug":"article-4-ai-literacy","label":"Article 4: AI literacy","summary":"Providers and deployers take measures that support the development of AI literacy.","legal_status":"applicable","deadline_at":"2025-02-02T00:00:00.000Z","human_page":"https://www.praxikon.com/en/verplichtingen/article-4-ai-literacy","api":"https://www.praxikon.com/api/v1/obligations?lang=en","official_source":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","citations":[{"kind":"official_fact","statement":"Since 27 July 2026, providers and deployers must take measures supporting the development of AI literacy. The provision does not require a guaranteed individual level.","source_id":"praxikon:eu:ai-act:source:reg-eu-2026-1744","source_locator":"Amendment of Article 4; entry into force 27 July 2026","source_url":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","eli":"http://data.europa.eu/eli/reg/2026/1744/oj"},{"kind":"editorial_interpretation","statement":"Evidence is primarily a proportionate record of measures by role and context, not one prescribed course or certificate.","source_id":"praxikon:eu:ai-act:source:commission-ai-literacy-qa","source_locator":"Questions on measures, formats, certificates and records","source_url":"https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers","eli":null},{"kind":"recommended_action","statement":"Inventory roles and AI systems, select appropriate measures and record the choice, implementation and periodic review.","source_id":"praxikon:eu:ai-act:source:commission-ai-literacy-qa","source_locator":"Implementation examples and evidence guidance","source_url":"https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers","eli":null}]},{"slug":"annex-iii-high-risk","label":"Annex III: high-risk AI","summary":"Classification route for standalone high-risk AI systems under Article 6(2) and Annex III.","legal_status":"upcoming","deadline_at":"2027-12-02T00:00:00.000Z","human_page":"https://www.praxikon.com/en/verplichtingen/annex-iii-high-risk","api":"https://www.praxikon.com/api/v1/obligations?lang=en","official_source":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","citations":[{"kind":"official_fact","statement":"The core rules in Chapter III, Sections 1 to 3, for systems under Article 6(2) and Annex III become applicable on 2 December 2027.","source_id":"praxikon:eu:ai-act:source:reg-eu-2026-1744","source_locator":"Amended Article 113, Article 6(2) and Annex III application date","source_url":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","eli":"http://data.europa.eu/eli/reg/2026/1744/oj"},{"kind":"editorial_interpretation","statement":"The later application date does not remove the classification question. An early classification record avoids design and procurement decisions without evidence.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 6 and Annex III","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"recommended_action","statement":"Document now the intended purpose, Annex III point, Article 6(3) assessment, profiling and selected registration path.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 6(2)-(4), Article 49 and Annex III","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"}]},{"slug":"article-14-human-oversight","label":"Article 14: human oversight","summary":"High-risk AI must be designed so that humans can effectively oversee it and intervene.","legal_status":"upcoming","deadline_at":"2027-12-02T00:00:00.000Z","human_page":"https://www.praxikon.com/en/verplichtingen/article-14-human-oversight","api":"https://www.praxikon.com/api/v1/obligations?lang=en","official_source":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","citations":[{"kind":"official_fact","statement":"Article 14 requires high-risk systems to be effectively overseeable by natural persons, with measures enabling them to understand the system, correctly interpret output, remain aware of automation bias, and decide not to use, to disregard or to stop the system.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 14(1)-(5)","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"official_fact","statement":"For the Annex III route this requirement applies from 2 December 2027; for high-risk AI in regulated products (Annex I) from 2 August 2028.","source_id":"praxikon:eu:ai-act:source:reg-eu-2026-1744","source_locator":"Amended Article 113 application dates","source_url":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","eli":"http://data.europa.eu/eli/reg/2026/1744/oj"},{"kind":"editorial_interpretation","statement":"Oversight on paper is not oversight: the law names automation bias explicitly, so a human who may only click through does not count. Effective oversight requires understanding, time and mandate, which ties directly into the Article 4 literacy measures.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 14(1)-(5)","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"},{"kind":"recommended_action","statement":"Appoint the overseeing persons per (upcoming) high-risk system now, train them specifically and record their mandate to intervene in writing.","source_id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","source_locator":"Article 14(1)-(5)","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj"}]}],"sources":[{"id":"praxikon:eu:ai-act:source:reg-eu-2024-1689","title":"EU Artificial Intelligence Act 2024/1689","publisher":"European Parliament and Council","canonical_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","eli":"http://data.europa.eu/eli/reg/2024/1689/oj","source_version":"original-oj-2024-07-12","verified_at":"2026-08-08T00:00:00.000Z"},{"id":"praxikon:eu:ai-act:source:reg-eu-2026-1744","title":"Digital Omnibus on AI 2026/1744","publisher":"European Parliament and Council","canonical_url":"https://eur-lex.europa.eu/eli/reg/2026/1744/oj","eli":"http://data.europa.eu/eli/reg/2026/1744/oj","source_version":"official-journal-2026-07-24","verified_at":"2026-08-08T00:00:00.000Z"},{"id":"praxikon:eu:ai-act:source:commission-ai-literacy-qa","title":"AI literacy questions and answers","publisher":"European Commission","canonical_url":"https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers","eli":null,"source_version":"updated-2026-07-27","verified_at":"2026-08-08T00:00:00.000Z"},{"id":"praxikon:eu:ai-act:source:commission-draft-high-risk-classification-guidelines","title":"Draft guidelines on the classification of high-risk AI systems","publisher":"European Commission (AI Office)","canonical_url":"https://digital-strategy.ec.europa.eu/en/library/draft-commission-guidelines-classification-high-risk-ai-systems","eli":null,"source_version":"draft-for-consultation-2026-05-19","verified_at":"2026-08-08T00:00:00.000Z"},{"id":"praxikon:eu:ai-act:source:cen-cenelec-jtc21","title":"CEN-CENELEC JTC 21: European standards under standardisation request M/613","publisher":"CEN-CENELEC JTC 21","canonical_url":"https://www.cencenelec.eu/areas-of-work/cen-cenelec-topics/artificial-intelligence/","eli":null,"source_version":"work-programme-checked-2026-08-08","verified_at":"2026-08-08T00:00:00.000Z"}],"first_actions":[{"label":"Submit the testing plan, obtain approval and register the test","summary":"Draw up a real-world testing plan, submit it to the market surveillance authority, obtain approval, register the test with a Union-wide unique single identification number, and record the division of roles with your deployer."},{"label":"Apply to a sandbox and agree the sandbox plan","summary":"Apply to the competent authority, agree a specific sandbox plan, and record which uncertainty about the Regulation you want resolved inside the sandbox."},{"label":"Take role- and context-specific AI literacy measures","summary":"Determine for each role, system and context which combination of instruction, guidance, practice or training is appropriate."}],"evidence":[{"label":"Dated and documented informed consent of test subjects","summary":"For every test subject you record freely given informed consent, covering five prescribed information elements, dated, documented, with a copy provided to the subject.","url":null},{"label":"Written proof of participation and the exit report","summary":"On request, the competent authority provides written proof of the activities successfully carried out, plus an exit report with results and learning outcomes. You can use that documentation in conformity assessment and in market surveillance.","url":null},{"label":"AI literacy measures record","summary":"Versioned record of roles, context, measures, participation or instruction and review moments.","url":null}],"guidance":[{"label":"No mandatory course format, no certificate, no exam and no AI officer","statement":"The Commission Q&A on AI literacy states that there is no one size fits all when it comes to AI literacy and that no strict requirements or mandatory trainings are imposed. On certification, the Q&A states literally that there is no need for a certificate and that organisations can keep an internal record of trainings or other guiding initiatives. On assessment, it states that Article 4 of the AI Act does not entail an obligation to measure the AI knowledge of employees. On governance, it states that no specific governance structure is mandated to comply with Article 4, so that unlike the data protection officer under the GDPR, no AI officer needs to be appointed. On the level, the Q&A states that following the Digital Omnibus amendment AI literacy remains an obligation for providers and deployers of AI systems, but that no specific or sufficient level is mandated and that the Regulation does not require guaranteeing any specific level of AI literacy of any individual. Against that, the Q&A states that simply relying on the AI systems' instructions for use or asking staff to read them might be ineffective, and that organisations should take into account general AI understanding within the organisation, whether they are a provider or a deployer, the risks associated with the systems deployed, staff knowledge gaps considering technical knowledge, experience, education and training, and contextual factors such as sector, purpose and affected populations. The Q&A further states that organisations may implement different levels of training or learning approaches depending on knowledge, experience, education and role, and that staff with a degree or experience in AI development are normally considered AI literate, while the organisation must still verify that those persons understand the specific AI systems of the organisation, know how to deal with them and are aware of all risks.","source_locator":"Commission Q&A on AI literacy, sections on required level, training formats, certificates, assessment of knowledge and governance structures (consulted 9 August 2026)"},{"label":"Article 4 reaches beyond your own staff, and the national supervisor enforces it","statement":"The Commission Q&A on AI literacy states that Article 4 applies to providers and deployers of AI systems and in addition to other persons dealing with the operation and use of AI systems on their behalf, covering persons broadly within the organisational remit, with a contractor, a service provider and a client given as examples. On clients, the Q&A states that they may need AI literacy depending on the specific risk, reasoning that affected persons should understand how decisions taken with the assistance of AI will have an impact on them. On geographic scope, the Q&A states that the AI Act's legal framework applies to both public and private actors inside and outside the EU as long as the AI system is placed on the Union market, used in the Union, or its use has an impact on people located in the EU. On supervision, the Q&A states that the supervision and enforcement of Article 4 is not with the AI Office but under the remit of national market surveillance authorities, and that supervision and enforcement began on 2 August 2026, while Article 4 itself entered into application on 2 February 2025. On sanctions, the Q&A states that national market surveillance authorities could impose penalties and other enforcement measures for infringements of Article 4, that this will be based on national laws that Member States were due to adopt by 2 August 2025, that any sanction must be proportionate and based on the individual case taking into account factors such as the nature and gravity of the infringement and its intentional or negligent character, and that sanctions are more likely if there is proof of an incident due to a lack of appropriate training and guidance. Article 4 is not listed in the enumeration in Article 99(4) of the AI Act, which covers only Articles 16, 22, 23, 24, 26, 31, 33(1), (3) and (4), 34 and 50, so the level of any penalty for Article 4 follows from national law rather than from the Regulation's own ceilings. The Q&A further states that Article 4 reinforces the transparency provisions of Article 13 and the human oversight provisions of Article 14 and indirectly contributes to the protection of affected persons, and that for deployers of high-risk systems the Article 26 obligation to ensure staff are trained to ensure human oversight is a distinct requirement; that requirement becomes applicable on 2 December 2027 for standalone Annex III systems and on 2 August 2028 for Annex I systems.","source_locator":"Commission Q&A on AI literacy, sections on target groups, geographic scope, supervision and enforcement, and sanctions (consulted 9 August 2026)"},{"label":"Article 6 has two separate routes to high-risk","statement":"The European Commission's draft guidelines on the classification of high-risk AI of 19 May 2026, which are expressly non-binding, state in paragraph (7) that an AI system is high-risk in two scenarios: first, where it is intended to be used as a safety component of a product, or is itself a product, covered by the Union harmonisation legislation listed in Annex I and required to undergo third-party conformity assessment; and second, where it falls within one of the use cases in the areas listed in Annex III. Paragraph (448) of those same draft guidelines notes that the Article 113 application dates have been postponed by the AI Omnibus to 2 December 2027 for the Article 6(2) route and 2 August 2028 for the Article 6(1) route.","source_locator":"Draft guidelines on high-risk AI classification (19 May 2026), General principles chapter, section II, paragraph (7); section V, paragraph (448)"},{"label":"Broadly positioned and general purpose AI systems: a disclaimer is not enough","statement":"According to the non-binding draft guidelines of 19 May 2026 on the classification of high-risk AI, paragraph (12) provides that where the instructions for use, contractual arrangements, terms of service, usage policy, promotional and sales materials or technical documentation present the AI system as broadly applicable across a generality of contexts and functions, and do not consistently limit its application or exclude high-risk uses, the system's intended purpose will be deemed to also encompass high-risk use cases and therefore qualify as high-risk. Under these draft guidelines this applies in particular where such uses are feasible and reasonably foreseeable given the system's functionalities and capabilities. The same paragraph states that merely asserting, for example in the terms of service, that high-risk uses are excluded is insufficient where the provider's overall presentation, examples or product positioning effectively provides for or promotes such uses, and that any limitations of use must be described clearly, concretely and coherently across all materials.","source_locator":"Draft guidelines on high-risk AI classification (19 May 2026), General principles chapter, section II.2, paragraph (12)"}],"examples":[{"label":"Candidate recommendation that automatically becomes a decision","situation":"An employer uses a system that ranks applicants and recommends a candidate to hire. In one setup a recruiter weighs that recommendation in their own assessment; in the other the outcome is applied automatically and a candidate is rejected without anyone looking at it.","outcome":"The Commission draft guidelines of 19 May 2026 address this case when determining whether an application falls under Annex III. The document is a consultation version: non-binding and not yet final.","lesson":"Assess a recruitment system on its intended purpose rather than on whether a recruiter reviews the output, because adding or removing human involvement does not change its high-risk classification.","source_locator":"Draft guidelines on high-risk AI classification, 19 May 2026, annex on Annex III","provenance":"official"},{"label":"Facial recognition at access control: the guard behind the camera counts too","situation":"An organisation secures the entrances to its buildings with facial recognition and uses that biometric access control to register visitors as well. When the system returns no match, a security officer reviews the camera images and decides personally whether someone may enter. The question is whose measures have to reach that officer: those of the supplier of the model, those of the department that deploys the system, or both.","outcome":"Article 4(1) provides that providers and deployers of AI systems shall take measures to support the development of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf. In doing so they must take into account technical knowledge, experience, education and training and the context the AI systems are to be used in, and consider the persons or groups of persons on whom the AI systems are to be used. The same provision states that this obligation does not require providers or deployers to guarantee any specific level of AI literacy of any individual.","lesson":"We read the phrase about the persons on whom the system is used as the centre of gravity for biometrics: whoever stands in front of the camera is subject to the outcome and has little to set against it. That argues for equipping the officer who decides for himself when no match comes back more substantively than the colleague who merely switches the system on and off. The article itself names no sufficient level and expressly states that you need not guarantee one, so where the floor lies for each role stays open. In our assessment a record kept per role, stating the choice made and the reason for it, is easier to defend than one organisation-wide session backed only by an attendance list.","source_locator":"Article 4(1)","provenance":"editorial"},{"label":"Police using AI in investigations: context sets how deep the training goes","situation":"A police force uses AI to search large volumes of investigation files and surface connections a detective would otherwise miss. The outputs feed into the choice of which suspect is pursued further and end up in documents that enter the criminal process. The question is whether one and the same basic instruction is enough for the analyst operating the model and for the detective who acts on its output.","outcome":"Article 4(1) requires providers and deployers of AI systems to take measures supporting the development of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf. The provision prescribes that they take into account technical knowledge, experience, education and training and the context the AI systems are to be used in, as well as the persons or groups of persons on whom the AI systems are to be used. It also states that this obligation does not require any specific level of AI literacy of any individual to be guaranteed.","lesson":"Article 4 requires you to weigh the context of use and the people the system is applied to, and in law enforcement both factors run high on our reading. Whether a general introduction to what AI can do is then enough for someone carrying an output into a file that affects a person's position as a suspect, we doubt, but the provision expressly names no level you must guarantee, so that floor is yours to justify. We would record for each role what someone must be able to recognise, for instance that a discovered connection is not yet evidence, and revisit that choice periodically.","source_locator":"Article 4(1)","provenance":"editorial"},{"label":"Newsroom with generative AI: do freelancers count within your measures?","situation":"A newsroom uses generative AI to prepare summaries, headlines and imagery, after which an editor finishes the piece and the desk decides to publish. Part of that work sits with freelancers, and an outside agency produces marketing content with the same tools. The question is whether your AI literacy measures must reach those freelancers and that agency, or only the people on the payroll.","outcome":"Article 4(1) is addressed to providers and deployers of AI systems and requires them to take measures supporting the development of AI literacy of their staff and of other persons dealing with the operation and use of AI systems on their behalf. The provision requires them to take into account technical knowledge, experience, education and training and the context the AI systems are to be used in, as well as the persons or groups of persons on whom the AI systems are to be used. It does not require any specific level of AI literacy of any individual to be guaranteed.","lesson":"Alongside staff, the text expressly names other persons dealing with the operation and use of AI systems on your behalf, and we read that as a functional boundary rather than a contractual one. On that reading a freelance editor using your tool inside your workflow and on your instruction sits within your measures, employment contract or not. The outside agency is a harder case: if it works in your environment and on your instruction, the argument that it acts on your behalf holds up, but if it runs its own tools on its own account it is a deployer in its own right, and Article 4 does not say your measures must cover that work. In practice, in our assessment, that means recording in your agreements who works in which role and what instruction you give, rather than trusting the other side to arrange it.","source_locator":"Article 4(1)","provenance":"editorial"}],"standards":[{"label":"prEN 18229-3: AI trustworthiness framework part 3, transparency and human oversight","summary":"The draft European standard for the transparency and oversight requirements of Articles 13 and 14 for high-risk AI systems.","statement":"prEN 18229-3 (AI trustworthiness framework, Part 3: Transparency and human oversight) is the JTC 21 deliverable under M/613 addressing Articles 13 and 14 of the AI Act: transparency and provision of information to deployers, and the design for effective human oversight of high-risk AI systems. As at June 2026 the deliverable was at the drafting stage. It has not yet been published as an EN and is not cited in the Official Journal. This deliverable concerns Article 14 (high-risk) and not the Article 50 transparency obligations, which apply since 2 August 2026."}],"definitions":[],"answer_page":"https://www.praxikon.com/en/antwoord/testomgeving-en-praktijktest","follow_up_questions":[{"question":"What does the Article 9 risk management system require from us?","url":"https://www.praxikon.com/en/antwoord/risicomanagement-hoog-risico"},{"question":"Do we have to keep logs of our AI system?","url":"https://www.praxikon.com/en/antwoord/logs-bewaren"},{"question":"How do we monitor our AI system after it goes live?","url":"https://www.praxikon.com/en/antwoord/monitoring-na-livegang"}],"disclaimer":"General interpretation, not legal advice. The official source remains authoritative.","methodology":"https://www.praxikon.com/en/methodologie"}