Preamble 91 to 100.
(91) Given the nature of AI systems and the risks to safety and fundamental rights possibly associated with their use, including as regards the need to ensure proper monitoring of the performance of an AI system in a real-life setting, it is appropriate to set specific responsibilities for deployers.
Deployers should in particular take appropriate technical and organisational measures to ensure they use high-risk AI systems in accordance with the instructions of use and certain other obligations should be provided for with regard to monitoring of the functioning of the AI systems and with regard to record-keeping, as appropriate.
Furthermore, deployers should ensure that the persons assigned to implement the instructions for use and human oversight as set out in this Regulation have the necessary competence, in particular an adequate level of AI literacy, training and authority to properly fulfil those tasks. Those obligations should be without prejudice to other deployer obligations in relation to high-risk AI systems under Union or national law.
(92) This Regulation is without prejudice to obligations for employers to inform or to inform and consult workers or their representatives under Union or national law and practice, including Directive 2002/14/EC of the European Parliament and of the Council, on decisions to put into service or use AI systems.
It remains necessary to ensure information of workers and their representatives on the planned deployment of high-risk AI systems at the workplace where the conditions for those information or information and consultation obligations in other legal instruments are not fulfilled.
Moreover, such information right is ancillary and necessary to the objective of protecting fundamental rights that underlies this Regulation. Therefore, an information requirement to that effect should be laid down in this Regulation, without affecting any existing rights of workers.
(93) Whilst risks related to AI systems can result from the way such systems are designed, risks can as well stem from how such AI systems are used. Deployers of high-risk AI system therefore play a critical role in ensuring that fundamental rights are protected, complementing the obligations of the provider when developing the AI system.
Deployers are best placed to understand how the high-risk AI system will be used concretely and can therefore identify potential significant risks that were not foreseen in the development phase, due to a more precise knowledge of the context of use, the persons or groups of persons likely to be affected, including vulnerable groups.
Deployers of high-risk AI systems listed in an annex to this Regulation also play a critical role in informing natural persons and should, when they make decisions or assist in making decisions related to natural persons, where applicable, inform the natural persons that they are subject to the use of the high-risk AI system.
This information should include the intended purpose and the type of decisions it makes. The deployer should also inform the natural persons about their right to an explanation provided under this Regulation. With regard to high-risk AI systems used for law enforcement purposes, that obligation should be implemented in accordance with Article 13 of Directive (EU) 2016/680.
(94) Any processing of biometric data involved in the use of AI systems for biometric identification for the purpose of law enforcement needs to comply with Article 10 of Directive (EU) 2016/680, that allows such processing only where strictly necessary, subject to appropriate safeguards for the rights and freedoms of the data subject, and where authorised by Union or Member State law.
Such use, when authorised, also needs to respect the principles laid down in Article 4 (1) of Directive (EU) 2016/680 including lawfulness, fairness and transparency, purpose limitation, accuracy and storage limitation.
(95) Without prejudice to applicable Union law, in particular Regulation (EU) 2016/679 and Directive (EU) 2016/680, considering the intrusive nature of post-remote biometric identification systems, the use of post-remote biometric identification systems should be subject to safeguards.
Post-remote biometric identification systems should always be used in a way that is proportionate, legitimate and strictly necessary, and thus targeted, in terms of the individuals to be identified, the location, temporal scope and based on a closed data set of legally acquired video footage.
In any case, post-remote biometric identification systems should not be used in the framework of law enforcement to lead to indiscriminate surveillance. The conditions for post-remote biometric identification should in any case not provide a basis to circumvent the conditions of the prohibition and strict exceptions for real time remote biometric identification.
(96) In order to efficiently ensure that fundamental rights are protected, deployers of high-risk AI systems that are bodies governed by public law, or private entities providing public services and deployers of certain high-risk AI systems listed in an annex to this Regulation, such as banking or insurance entities, should carry out a fundamental rights impact assessment prior to putting it into use. Services important for individuals that are of public nature may also be provided by private entities. Private entities providing such public services are linked to tasks in the public interest such as in the areas of education, healthcare, social services, housing, administration of justice.
The aim of the fundamental rights impact assessment is for the deployer to identify the specific risks to the rights of individuals or groups of individuals likely to be affected, identify measures to be taken in the case of a materialisation of those risks. The impact assessment should be performed prior to deploying the high-risk AI system, and should be updated when the deployer considers that any of the relevant factors have changed.
The impact assessment should identify the deployer’s relevant processes in which the high-risk AI system will be used in line with its intended purpose, and should include a description of the period of time and frequency in which the system is intended to be used as well as of specific categories of natural persons and groups who are likely to be affected in the specific context of use.
The assessment should also include the identification of specific risks of harm likely to have an impact on the fundamental rights of those persons or groups. While performing this assessment, the deployer should take into account information relevant to a proper assessment of the impact, including but not limited to the information given by the provider of the high-risk AI system in the instructions for use.
In light of the risks identified, deployers should determine measures to be taken in the case of a materialisation of those risks, including for example governance arrangements in that specific context of use, such as arrangements for human oversight according to the instructions of use or, complaint handling and redress procedures, as they could be instrumental in mitigating risks to fundamental rights in concrete use-cases.
After performing that impact assessment, the deployer should notify the relevant market surveillance authority. Where appropriate, to collect relevant information necessary to perform the impact assessment, deployers of high-risk AI system, in particular when AI systems are used in the public sector, could involve relevant stakeholders, including the representatives of groups of persons likely to be affected by the AI system, independent experts, and civil society organisations in conducting such impact assessments and designing measures to be taken in the case of materialisation of the risks. The European Artificial Intelligence Office (AI Office) should develop a template for a questionnaire in order to facilitate compliance and reduce the administrative burden for deployers.
(97) The notion of general-purpose AI models should be clearly defined and set apart from the notion of AI systems to enable legal certainty. The definition should be based on the key functional characteristics of a general-purpose AI model, in particular the generality and the capability to competently perform a wide range of distinct tasks.
These models are typically trained on large amounts of data, through various methods, such as self-supervised, unsupervised or reinforcement learning. General-purpose AI models may be placed on the market in various ways, including through libraries, application programming interfaces (APIs), as direct download, or as physical copy. These models may be further modified or fine-tuned into new models.
Although AI models are essential components of AI systems, they do not constitute AI systems on their own. AI models require the addition of further components, such as for example a user interface, to become AI systems. AI models are typically integrated into and form part of AI systems.
This Regulation provides specific rules for general-purpose AI models and for general-purpose AI models that pose systemic risks, which should apply also when these models are integrated or form part of an AI system. It should be understood that the obligations for the providers of general-purpose AI models should apply once the general-purpose AI models are placed on the market.
When the provider of a general-purpose AI model integrates an own model into its own AI system that is made available on the market or put into service, that model should be considered to be placed on the market and, therefore, the obligations in this Regulation for models should continue to apply in addition to those for AI systems.
The obligations laid down for models should in any case not apply when an own model is used for purely internal processes that are not essential for providing a product or a service to third parties and the rights of natural persons are not affected. Considering their potential significantly negative effects, the general-purpose AI models with systemic risk should always be subject to the relevant obligations under this Regulation.
The definition should not cover AI models used before their placing on the market for the sole purpose of research, development and prototyping activities. This is without prejudice to the obligation to comply with this Regulation when, following such activities, a model is placed on the market.
(98) Whereas the generality of a model could, inter alia, also be determined by a number of parameters, models with at least a billion of parameters and trained with a large amount of data using self-supervision at scale should be considered to display significant generality and to competently perform a wide range of distinctive tasks.
(99) Large generative AI models are a typical example for a general-purpose AI model, given that they allow for flexible generation of content, such as in the form of text, audio, images or video, that can readily accommodate a wide range of distinctive tasks.
(100) When a general-purpose AI model is integrated into or forms part of an AI system, this system should be considered to be general-purpose AI system when, due to this integration, this system has the capability to serve a variety of purposes. A general-purpose AI system can be used directly, or it may be integrated into other AI systems.
Note: This is the final text of the Artificial Intelligence Act (Regulation (EU) 2024/1689), as it was published in the Official Journal of the European Union in July, 2024.