Article 76: Supervision of Testing in Real World Conditions by Market Surveillance Authorities
1.
Market surveillance authorities shall have competences and powers to ensure that testing in real world conditions is in accordance with this Regulation.
Where testing in real world conditions is based on Article 60a, any reference to a market surveillance authoritymarket surveillance authorityAny national authority responsible for carrying out market surveillance on the territory of the Member State concernedRegulation (EU) 2019/1020 Art. 3(4) in this Article shall be construed as a reference to the national competent authoritynational competent authoritymeans a notifying authoritynotifying authoritymeans the national authority responsible for setting up and carrying out the necessary procedures for the assessment, designation and notification of conformity assessment bodies and for their monitoringArticle 3(19) or a market surveillance authority; as regards AI systems put into service or used by Union institutions, agencies, offices and bodies, references to national competent authorities or market surveillance authorities in this Regulation shall be construed as references to the European Data Protection SupervisorArticle 3(48) or appropriate authority under the Union harmonisation legislation listed in Section B of Annex I, and references to Article 60 shall be construed as references to Article 60a, as appropriate.Not present before the amendment.
2.
Where testing in real world conditions is conducted for AI systems that are supervised within an AI regulatory sandboxAI regulatory sandboxmeans a controlled framework set up by a competent authority which offers providers or prospective providers of AI systems the possibility to develop, train, validate and test, where appropriate in real-world conditions, an innovative AI systemAI systemmeans a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environmentsArticle 3(1), pursuant to a sandbox plansandbox planmeans a document agreed between the participating providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) and the competent authority describing the objectives, conditions, timeframe, methodology and requirements for the activities carried out within the sandboxArticle 3(54) for a limited time under regulatory supervisionArticle 3(55) under Article 58, the market surveillance authorities shall verify the compliance with Article 60 as part of their supervisory role for the AI regulatory sandboxAI regulatory sandboxmeans a controlled framework set up by a competent authority which offers providers or prospective providers of AI systems the possibility to develop, train, validate and test, where appropriate in real-world conditions, an innovative AI systemAI systemmeans a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environmentsArticle 3(1), pursuant to a sandbox plansandbox planmeans a document agreed between the participating providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) and the competent authority describing the objectives, conditions, timeframe, methodology and requirements for the activities carried out within the sandboxArticle 3(54) for a limited time under regulatory supervisionArticle 3(55). Those authorities may, as appropriate, allow the testing in real world conditions to be conducted by the providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) or prospective providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3), in derogation from the conditions set out in Article 60(4), points (f) and (g).
3.
Where a market surveillance authoritymarket surveillance authorityAny national authority responsible for carrying out market surveillance on the territory of the Member State concernedRegulation (EU) 2019/1020 Art. 3(4) has been informed by the prospective providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3), the providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) or any third party of a serious incidentserious incidentmeans an incident or malfunctioning of an AI systemAI systemmeans a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environmentsArticle 3(1) that directly or indirectly leads to any of the following:Article 3(49) or has other grounds for considering that the conditions set out in Articles 60 and 61 are not met, it may take either of the following decisions on its territory, as appropriate:
(a)
to suspend or terminate the testing in real world conditions;
(b)
to require the providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) or prospective providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) and the deployerdeployermeans a natural or legal person, public authority, agency or other body using an AI system under its authority except where the AI system is used in the course of a personal non-professional activityArticle 3(4) or prospective deployerdeployermeans a natural or legal person, public authority, agency or other body using an AI system under its authority except where the AI system is used in the course of a personal non-professional activityArticle 3(4) to modify any aspect of the testing in real world conditions.
4.
Where a market surveillance authoritymarket surveillance authorityAny national authority responsible for carrying out market surveillance on the territory of the Member State concernedRegulation (EU) 2019/1020 Art. 3(4) has taken a decision referred to in paragraph 3 of this Article, or has issued an objection within the meaning of Article 60(4), point (b), the decision or the objection shall indicate the grounds therefor and how the providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) or prospective providerprovidermeans a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of chargeArticle 3(3) can challenge the decision or objection.
5.
Where applicable, where a market surveillance authoritymarket surveillance authorityAny national authority responsible for carrying out market surveillance on the territory of the Member State concernedRegulation (EU) 2019/1020 Art. 3(4) has taken a decision referred to in paragraph 3, it shall communicate the grounds therefor to the market surveillance authorities of other Member States in which the AI systemAI systemmeans a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environmentsArticle 3(1) has been tested in accordance with the testing plan.