When must a serious incident be reported under Article 73 of the EU AI Act?

A provider of a high-risk AI system placed on the Union market reports a serious incident not later than 15 days after it or, where applicable, the deployer becomes aware of it, 10 days for a death and two days for a widespread infringement or a serious and irreversible disruption of critical infrastructure (Article 73(2) to (4)). Each is an outer limit: each of those paragraphs also says "immediately".

Written and last checked 9 October 2026 against Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744. Article 73 was not amended; the amending act inserted Article 75(1a), which sends some of these reports to the AI Office.

What counts as a serious incident?

Article 3, point (49) defines it as "an incident or malfunctioning of an AI system that directly or indirectly leads to any of the following": "the death of a person, or serious harm to a person's health" (point (a)); "a serious and irreversible disruption of the management or operation of critical infrastructure" (point (b)), with critical infrastructure as defined in Article 2, point (4), of Directive (EU) 2022/2557 (Article 3, point (62)); "the infringement of obligations under Union law intended to protect fundamental rights" (point (c)); and "serious harm to property or the environment" (point (d)). Points (c) and (d) need no injury to anyone.

Article 73(3) also names the widespread infringement, which Article 3, point (61) defines as "any act or omission contrary to Union law protecting the interest of individuals", in two points. Under point (a), it "has harmed or is likely to harm the collective interests of individuals residing in at least two Member States other than the Member State in which: (i) the act or omission originated or took place; (ii) the provider concerned, or, where applicable, its authorised representative is located or established; or (iii) the deployer is established, when the infringement is committed by the deployer". Under point (b), it "has caused, causes or is likely to cause harm to the collective interests of individuals and has common features, including the same unlawful practice or the same interest being infringed, and is occurring concurrently, committed by the same operator, in at least three Member States".

How many days does each limit allow, and what starts it?

Under Article 73(2), the report "shall be made immediately after the provider has established a causal link between the AI system and the serious incident or the reasonable likelihood of such a link, and, in any event, not later than 15 days after the provider or, where applicable, the deployer, becomes aware of the serious incident." Its second subparagraph adds that the period "shall take account of the severity of the serious incident".

Under Article 73(3), "in the event of a widespread infringement or a serious incident as defined in Article 3, point (49)(b)", the report "shall be provided immediately, and not later than two days after the provider or, where applicable, the deployer becomes aware of that incident." Article 73(3) names no causal link: the two days run from awareness.

Under Article 73(4), "in the event of the death of a person", the report "shall be provided immediately after the provider or the deployer has established, or as soon as it suspects, a causal relationship between the high-risk AI system and the serious incident, but not later than 10 days after the date on which the provider or, where applicable, the deployer becomes aware of the serious incident." Outside the narrower routes in Article 73(9) and (10) below, serious harm to health, the infringement of fundamental-rights obligations and serious harm to property or the environment keep the 15 days of Article 73(2), unless they also amount to a widespread infringement, which takes the two days of Article 73(3).

For all three, Article 73(5) provides: "Where necessary to ensure timely reporting, the provider or, where applicable, the deployer, may submit an initial report that is incomplete, followed by a complete report."

Who reports, and to whom?

The provider: Article 73(1) has "Providers of high-risk AI systems placed on the Union market" report "to the market surveillance authorities of the Member States where that incident occurred". Whether you are the provider is the first thing to settle.

Regulation (EU) 2026/1744 added a derogation. Under Article 75(1a), providers of high-risk systems within the AI Office's exclusive competence under Article 75(1), as replaced, report serious incidents to the AI Office, with Article 73(2) to (9) applying mutatis mutandis; the AI Office promptly transmits the relevant information to the market surveillance authority of the Member State where the provider or its legal representative is situated. Article 75(1), first subparagraph, names two classes. Point (a) is systems based on general-purpose AI models where the model and the system are developed by the same provider or within the same undertaking, "with the exception of: (i) AI systems related to products covered by the Union harmonisation legislation listed in Annex I; (ii) AI systems referred to in point 2 of Annex III; (iii) AI systems provided by law enforcement authorities, border management authorities and financial institutions, insofar as those AI systems fall under Article 74(6); and (iv) AI systems referred to in point 8 of Annex III as regards the administration of justice". Point (b) is systems that constitute or are integrated into a very large online platform or very large online search engine designated under Regulation (EU) 2022/2065.

A deployer that has identified a serious incident shall "immediately inform first the provider, and then the importer or distributor and the relevant market surveillance authorities of that incident"; if it cannot reach the provider, Article 73 applies mutatis mutandis (Article 26(5), first subparagraph). That does not cover sensitive operational data of deployers that are law enforcement authorities. For deployers that are financial institutions "subject to requirements regarding their internal governance, arrangements or processes under Union financial services law", the second subparagraph deems "the monitoring obligation set out in the first subparagraph" fulfilled by complying with those rules; whether that reaches the incident duty, the text does not say.

Three narrower routes. Under Article 73(9), for Annex III systems placed on the market or put into service by providers subject to Union legislative instruments laying down reporting obligations equivalent to those of the Regulation, notification is limited to incidents under Article 3, point (49)(c). Article 73(10) sets the same limit for systems that are safety components of devices, or are themselves devices, covered by Regulations (EU) 2017/745 and (EU) 2017/746, and sends the notification "to the national competent authority chosen for that purpose by the Member States where the incident occurred". Article 60(7) sends a serious incident identified in testing in real world conditions to the national market surveillance authority in accordance with Article 73, and has the provider or prospective provider adopt immediate mitigation measures or, failing that, suspend the testing until mitigation takes place or terminate it, with a procedure for the prompt recall of the system upon termination.

What must the provider do after reporting?

Article 73(6) has the provider "without delay, perform the necessary investigations in relation to the serious incident and the AI system concerned", including "a risk assessment of the incident, and corrective action". During the investigations it cooperates with the competent authorities and, where relevant, with the notified body concerned. It "shall not perform any investigation which involves altering the AI system concerned in a way which may affect any subsequent evaluation of the causes of the incident, prior to informing the competent authorities of such action."

On the authority side, the market surveillance authority, on receiving a notification of a point (49)(c) incident, informs the national public authorities or bodies referred to in Article 77(1), as replaced (Article 73(7)); the market surveillance authority takes appropriate measures under Article 19 of Regulation (EU) 2019/1020 within seven days from receiving the notification and follows that Regulation's notification procedures (Article 73(8)); and national competent authorities notify the Commission immediately of any serious incident, whether or not they have acted on it, in accordance with Article 20 of Regulation (EU) 2019/1020 (Article 73(11)).

Article 73(7) also required the Commission to issue dedicated guidance on the paragraph 1 obligations by 2 August 2025, to be assessed regularly; that guidance is not among the sources this guide was checked against.

When does Article 73 apply?

Article 73 is in Chapter IX, Section 2, which no point in the third paragraph of Article 113 covers, so it takes the general date in the second paragraph: 2 August 2026. The duty attaches to providers of high-risk AI systems, and the classification rules in Chapter III, Section 1 apply from 2 December 2027 for systems classified as high-risk under Article 6(2) and Annex III, and from 2 August 2028 under Article 6(1) and Annex I (Article 113, third paragraph, point (c), as replaced by Regulation (EU) 2026/1744). The deployer's duty in Article 26(5) and the quality management system's "procedures related to the reporting of a serious incident in accordance with Article 73" (Article 17(1), point (i)) sit in Section 3 and take those later dates. The amending regulation does not say how Article 73 operates between 2 August 2026 and the date the classification rules apply to a system. That is an open question, and this guide does not resolve it.

For a high-risk system placed on the market or put into service before the date of application of Chapter III, other than the Annex X systems in Article 111(1), Article 111(2), as replaced, applies the Regulation, Article 5 aside, only if, as from that date, the system is subject to significant changes in its design; providers and deployers of high-risk systems intended to be used by public authorities take the necessary steps to comply by 2 August 2030. Article 111(2) names one date where Article 113 now gives two; systems already on the market has the detail.

What does Complipath do here, and what does it not?

Serious incident reporting (Article 73): we found no support for this in what we have built. Post-market monitoring (Article 72) has the same answer in what Complipath is. The AI inventory records each system with its classification and the article behind it, so the high-risk systems Article 73 reaches are visible. Each duty that follows carries an owner, a status and a date. Evidence is linked with the passage and its page, and a change to a provision your records cite is emailed per affected system.

What this means for you

If you're a provider: write the procedure before you need it: who decides that an incident is serious, which limit applies, who sends the initial report to which authority and who tells the competent authorities before an investigation alters the system in a way that may affect the evaluation of causes. Article 17(1), point (i) requires procedures for reporting serious incidents in your quality management system from your Chapter III date.

If you're a deployer: inform the provider immediately, then the importer or distributor and the market surveillance authorities (Article 26(5)), and put the provider's incident contact in the contract: if you cannot reach the provider, Article 73 applies mutatis mutandis. Article 26 in full has the rest.

FAQ

Is a serious incident always about someone getting hurt? No. Article 3, point (49) covers an incident or malfunctioning that directly or indirectly leads to a death or serious harm to health, a serious and irreversible disruption of critical infrastructure, the infringement of obligations under Union law intended to protect fundamental rights or serious harm to property or the environment. The last two need no injury to anyone.

Can we report before we know what caused it? Yes. Where necessary to ensure timely reporting, Article 73(5) allows an initial report that is incomplete, followed by a complete report. For a death, the report is due as soon as a causal relationship is suspected, and not later than 10 days after awareness (Article 73(4)). Article 73(3) names no causal link at all.

Does a deployer ever report to the authority? Yes. A deployer that has identified a serious incident immediately informs first the provider, and then the importer or distributor and the relevant market surveillance authorities (Article 26(5)). If it cannot reach the provider, Article 73 applies mutatis mutandis. The obligation does not cover sensitive operational data of deployers that are law enforcement authorities.

Does Article 73 apply to general-purpose AI models? Not to the model as such: Article 73 binds providers of high-risk AI systems. A provider of a general-purpose AI model with systemic risk keeps track of, documents and reports relevant information about serious incidents and possible corrective measures, without undue delay, to the AI Office and, as appropriate, to national competent authorities (Article 55(1), point (c)).

Sources: Regulation (EU) 2024/1689 (EUR-Lex), Article 3, points (49), (61) and (62), Article 6, Article 17(1), point (i), Article 26(5), Article 55(1), point (c), Article 60(7), Article 72, Article 73, Article 74(6), Article 75(1), Article 77(1), Article 111 and Article 113; as amended by Regulation (EU) 2026/1744 (EUR-Lex), which inserted Article 75(1a) and replaced Article 75(1), Article 77(1), Article 111(2) and Article 113, third paragraph, point (c). Article 73 was not amended.

Where this question meets the product: what counts as evidence for each requirement. Answering it in a customer's questionnaire: how to answer AI questions in a supplier questionnaire.