bc-ontologyGovernments, institutions, geography and law
Pre-release: this release has not been published. Identifiers are permanent; descriptions may still change.

source excerpt from https://enterprise.gov.ie/en/legislation/legislation-files/general-scheme-of-the-regulation-of-artificial-intelligence-bill-2026.pdf

https://ontology.bommaritollc.com/id/01a135e4-57d9-73a8-9de5-6ea6a3e124db
oa:SpecificResource
Download TurtleJSON-LDN-TriplesMarkdown
Cite this
source excerpt from https://enterprise.gov.ie/en/legislation/legislation-files/general-scheme-of-the-regulation-of-artificial-intelligence-bill-2026.pdf. bc-ontology, release 0.1.0-rc38, 2026-10-21. https://ontology.bommaritollc.com/id/01a135e4-57d9-73a8-9de5-6ea6a3e124db

Elsewhere

Facts

adms:identifier
oa:bodyValue
{"legal_state":"unknown","passage":{"end_byte":155366,"instrument_iri":"https://ontology.bommaritollc.com/id/01a1351f-f68e-77d8-8511-7755bbd6fb10","original_sha256":"ca0f1c74524fccb3d57393b26eb5ba44455fc33c9c1330b3b52caf002f14763d","parsed_text_sha256":"4e6f575b2c6aaadc72a511c5bf5206a6272d9b5b31f0ee9359f58cbf5f2ba4c9","source_url":"https://enterprise.gov.ie/en/legislation/legislation-files/general-scheme-of-the-regulation-of-artificial-intelligence-bill-2026.pdf","start_byte":143428},"policy":"retained-legal-source-passage-v1","quote":"and develop technical systems to facilitate consistent and streamlined reporting by MSAs. This\n\nsupports a coordinated national approach to reporting and ensures that Ireland meets its obligations\n\nunder EU law.\n\nSubhead (3) sets out that MSAs must actively participate in the Cooperation Forum established by\n\nthe AI Office of Ireland. The Forum will serve as a platform for communication and cooperation\n\namong competent authorities, supporting the consistent application of the AI Act, sharing of best\n\npractices, and effective supervision and enforcement at national level. This is provided for in Head\n\n34.\n\nPage 66 of 180\n\n\nHead 42 - Serious incidents\n\n(1) Relevant Market Surveillance Authorities will receive reports of serious incidents from\n\nproviders of High-Risk AI (HRAI) systems where the incident occurred.\n\n(2) Procedures to manage the reporting of serious incidents w will be carried out in accordance with\n\nArticle 19 of the Market Surveillance Regulation or union harmonisation legislation.\n\n(3) The relevant Market Surveillance Authority t that received the relevant report will notify public\n\nauthorities, the Commission, the AI Office of Ireland or bodies designated under Article 77(1).\n\n(4) The relevant Market Surveillance Authorities for high-risk AI systems in Annex III AI Act, that\n\nare already subject to equivalent Union reporting obligations the notifications are limited to\n\nserious incident reports specific only to Article 3, point (49)(c) from providers operating in\n\nthese respective subject domains.\n\n(5) The Market Surveillance Authority designated in regulation 3 of SI No. 366 of 2025 w will receive\n\nreports of serious incidents in relation to medical devices and safety components of devices\n\ncovered by Regulations (EU) 2017/745 and (EU) 2017/746.\n\nExplanatory Note:\n\n\nIncident reporting is required under Article 73 and Article 76 of the AI Act. This Head sets out that\n\nthe relevant Market Surveillance Authorities will ensure mechanisms are in place for providers to\n\nreport serious incidents, and that the procedures outlined in the Market Surveillance Regulation\n\n2019/1020 are followed, as well as notifying public authorities, the Commission and Office. Article\n\n73 (7) of the AI Act states that the Commission will develop dedicated guidelines to facilitate\n\ncompliance with the reporting obligation in Article 73 (1). Draft guidelines were published by the\n\nEuropean Commission on 26 September 20255for stakeholder and public consultation and when\n\n,\n\nfinalised they will provide further information and clarification for this provision.\n\nProviders must notify all serious incidents to the market surveillance authorities of the Member\n\nStates where the incident occurred. The notification must be made immediately after the provider\n\nhas established a causal link between the AI-system and between the serious incident or the\n\nreasonable likelihood of such a connection, and in each case no later than 15 days after the serious\n\nincident has come to the attention of the provider or, as the case may be, the deployer.\n\nIn the case of a fatality, the notification must be made no later than 10 days after the date on which\n\nthe serious incident has come to the attention of the provider or, as the case may be, the deployer.\n\nHowever, in the event of a ’widespread incident or a serious incident as defined in Article 3(49)(b)\n\n((b) a serious and irreversible disruption of the management and operation of critical infrastructure,\n\nthe report must be provided immediately and no later than 2 days after the provider or deployer\n\nbecomes aware of the incident.\n\nThe provider must immediately carry out the necessary investigations related to the serious incident\n\nand the artificial intelligence system in question. This must include a risk assessment of the incident\n\nand corrective measures.\n\n5 AI Act: Commission issues draft guidance and reporting template on serious AI incidents, and seeks\n\nstakeholders' feedback | Shaping Europe’s digital future\n\nPage 67 of 180\n\n\nArticle 3 (49) of the AI Act defines serious incident as follows:\n\nany incident or malfunctioning of an AI system that directly or indirectly leads to any of the\n\nfollowing:\n\n(a) the death of a person or serious damage to a person’s health;\n\n(b) a serious and irreversible disruption of the management and operation of critical\n\ninfrastructure;\n\n(c) the infringement of obligations under Union law intended to protect fundamental rights.”\n\nSubhead (1) Market: Surveillance Authorities must ensure mechanisms are in place for providers to\n\n\nreport serious incidents, as defined in Article 3(49) of the AI Act—covering harm to health,\n\ndisruption of critical infrastructure, or breaches of fundamental rights.\n\nSubhead (2) provides a procedural reference to Article 19 of Regulation 2019/1020 which outlines\n\n\nthe procedures MSAs must follow to manage incident reports, including investigation, coordination,\n\nand enforcement actions. It also provides for the sector specific harmonised legislation. E.g. for\n\nmedical devices there are separate definitions, processes and systems in place to receive reports of\n\nincidents, to investigate and to report incidents on separate systems.\n\nSubhead (3) is about the actions that need to be taken upon receiving a report, where an MSA must\n\n\ninform the relevant national public authorities or bodies listed in Article 77(1) of the AI Act, ensuring\n\ncoordinated response and oversight. It is provided for in Article 73 (7) of the AI Act that the\n\nCommission will develop dedicated guidelines to facilitate compliance with the reporting obligation\n\nin Article 73 (1). The guidelines have not been published yet, but they will provide further\n\ninformation and clarification for this provision.\n\nSubhead (4) specifies that for high-risk AI systems under Annex III that are already subject to\n\n\nequivalent Union reporting obligations, notification is limited to incidents involving breaches of\n\nfundamental rights (Article 3(49)(c)).\n\nSubhead (5) is a specific requirement of the AI Act, and addresses the regulator of medical devices,\n\n\nThe Health Products Regulatory Authority (HPRA), which will handle serious incident reports related\n\nto AI medical devices or safety components of medical devices governed by Regulations (EU)\n\n2017/745 and 2017/746.\n\nPage 68 of 180\n\n\nHead 43 - Temporary authorisation of High-Risk AI Systems for exceptional\n\ncircumstances\n\n(1) A relevant Market Surveillance Authority may, upon receiving a duly justified request in a\n\nformat from a provider, authorise the placing on the market or putting into service of a high\u0002\n\nrisk AI system without prior completion of the conformity assessment procedures where\n\nsuch authorisation is necessary to address exceptional circumstances.\n\n(2) A derogation may be granted only where:\n\n(a) the high-risk AI system is required to respond to an urgent public need, including\n\nbut not limited to public health, safety, or national security;\n\n(b) the full conformity assessment cannot be completed in time to meet that need;\n\nand\n\n(c) the provider demonstrates that the system complies with applicable safety and\n\nfundamental rights protections to the extent feasible.\n\n(3) Any authorisation granted under this section shall:\n\n(a) be limited in scope and duration to what is strictly necessary;\n\n(b) specify the conditions under which the system may be used;\n\n(c) require ongoing monitoring and reporting by the provider to the Market\n\nSurveillance Authority.\n\n(4) The relevant Market Surveillance Authority shall:\n\n(a) notify the AI Office of Ireland of any derogation granted under this section;\n\n(b) publish a summary of the derogation decision, including its justification and\n\nconditions, unless doing so would compromise public safety or national security.\n\n(5) The Market Surveillance Authority may review and, where appropriate, revoke the\n\nderogation if:\n\n(a) the conditions under which it was granted are no longer met;\n\n(b) the system is found to pose a risk to health, safety, or fundamental rights.\n\nExplanatory Note:\n\n\nThis Head provides that a Market Surveillance Authority may authorise the placing on the market or\n\nthe putting into service of specific high-risk AI systems within the State, for exceptional reasons of\n\npublic security or the protection of life and health of persons, environmental protection or the\n\nprotection of key industrial and infrastructural assets (Article 43 derogation), as a time-specific\n\nmeasure.\n\nArticle 43 of the AI Act sets out the rules for conformity assessment—the process by which providers\n\nof high-risk AI systems must prove that their systems meet the legal requirements before they can\n\nbe placed on the EU market. The conformity assessment is a formal check to ensure that a high-risk\n\nsystem complies with the rules in Chapter III, Section 2.\n\nPage 69 of 180\n\n\nThe derogation from Article 43 is provided for and set out in detail in Article 46 of the AI Act. It\n\nallows a Market Surveillance Authority (MSA) to authorise the placing on the market of a high-risk AI\n\nsystem without a full conformity assessment, but only in exceptional cases.\n\nConditions for Derogation:\n\n• The provider must submit a duly justified request.\n\n• The derogation is granted only when necessary, for example:\n\no in response to urgent public needs;\n\no for testing in real-world conditions under supervision;\n\no when the full assessment is not feasible in time, but the system is needed (e.g.\n\nduring a public emergency).\n\nSuch a derogation might be used in a scenario where an AI system is used for emergency response\n\nduring a natural disaster and there is not time for full conformity assessment. It would be in very\n\nexceptional circumstances and for a defined time.\n\nPage 70 of 180\n\n\nHead 44 - Request to organise testing of a high-risk AI system\n\n(1) A relevant Market Surveillance Authority shall, upon receiving a reasoned request from a\n\nFundamental Rights Authority identified pursuant to Article 77 of the AI Act take all necessary\n\nsteps to assess the high-risk AI system concerned, including but not limited to:\n\na) requesting and reviewing relevant documentation from the provider or user of the AI\n\nsystem;\n\nb) conducting or commissioning technical testing of the AI system;\n\nc) engaging with the Fundamental Rights Authority to clarify the nature of the concerns\n\nraised.\n\n(2) A request under subsection (1) shall be made in writing and shall include a statement of the\n\ngrounds upon which the Fundamental Rights Authority considers that the AI system may present\n\na risk to fundamental rights.\n\n(3) The relevant Market Surveillance Authority shall respond to the requesting authority within a\n\nreasonable time following receipt of the request indicating:\n\na. the actions it proposes to take; or\n\nb. the reasons why no action is considered necessary.\n\n(4) Where the relevant Market Surveillance Authority determines that testing is required, it shall\n\nensure that such testing is carried out in a timely manner and that the results are communicated\n\nto the requesting authority.\n\n(5) The Minister may issue guidelines or regulations specifying procedures for the making, handling,\n\nand responding to requests under this section.\n\nExplanatory Note\n\n\nThis Head sets out how a relevant Market Surveillance Authority should fulfil its obligation in Article\n\n77 (3) of the AI Act which means that a Fundamental Rights body ‘may make a reasoned request to\n\nthe MSA to organise testing of the high-risk AI System through technical means’.\n\nArticle 77 is further explained in an interpretative note produced by the AI Office in October 2024,\n\nwhich says:\n\nArticle 77 of the AI Act seeks to address the situation in which, in a case of misuse or\n\nmalfunctioning of high-risk AI systems falling under Annex III, the result is a fundamental rights\n\nviolation. Therefore, national authorities or bodies that are responsible to enforce or supervise\n\n","statutory_number":null,"whole_instrument_recall":"not_established"}rdf:JSON
oa:hasSource
prov:wasDerivedFrom
schema:isPartOf

Related

Errata in its text (3)oa:hasTarget