AI Literacy - Questions & Answers

Article 4 of the AI Act requires that providers and deployers of AI systems should take measures to ensure a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf. They are obliged to do so by taking into account their technical knowledge, experience, education and training of the staff and other persons as well as the context the AI systems are to be used in and the persons on whom the AI systems are to be used.

The concept of AI literacy mentioned in article 4 of the AI Act relies on the definition of the term given in article 3(56) of the AI Act, according to which: ‘AI literacy’ means skills, knowledge and understanding that allow providers, deployers and affected persons, taking into account their respective rights and obligations in the context of this Regulation, to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause. 

Article 4 of the AI Act is a key provision to ensure that all providers and deployers of AI systems equip their staff with the right skills, knowledge and understanding of the system(s) provided or deployed. This concerns anyone in the organisation directly dealing with an AI system and reinforces the provisions of transparency (Article 13 of the AI Act) and human oversight (Article 14 of the AI Act) included in the Regulation. At the same time, Article 4 indirectly contributes to the protection of affected persons, because it ensures an effective application of the AI Act rules.  

“Persons dealing with the operation and use of AI systems on behalf of providers/deployers” means that these are not employees, but persons broadly under the organisational remit. It could be, for example, a contractor, a service provider, a client. 

Article 3 (56) mentions affected persons. Also recital 20 reads: ‘In order to obtain the greatest benefits from AI systems while protecting fundamental rights, health and safety and to enable democratic control, AI literacy should equip providers, deployers and affected persons with the necessary notions to make informed decisions regarding AI systems. Those notions may vary with regard to the relevant context and can include understanding the correct application of technical elements during the AI system’s development phase, the measures to be applied during its use, the suitable ways in which to interpret the AI system’s output, and, in the case of affected persons, the knowledge necessary to understand how decisions taken with the assistance of AI will have an impact on them. In the context of the application this Regulation, AI literacy should provide all relevant actors in the AI value chain with the insights required to ensure the appropriate compliance and its correct enforcement. Furthermore, the wide implementation of AI literacy measures and the introduction of appropriate follow-up actions could contribute to improving working conditions and ultimately sustain the consolidation, and innovation path of trustworthy AI in the Union’. Given the above, dependent on the specific risk, it can be useful to have literacy as well for customers/clients.”

Article 4 of the AI Act does not entail an obligation to measure the knowledge of AI of employees. Yet, it affirms that AI providers and deployers should ensure a sufficient level of AI literacy taking into account the technical knowledge, experience, education and training of employees.

Yes, the AI Act distinguishes between AI models (GPAI models), AI systems, including GPAI systems, prohibited, high-risk systems. For more details, please also see the guidelines on AI system definition, published by the Commission on 6 February to assist providers and other relevant persons in determining whether a software system constitutes an AI system and facilitate the effective application of the rules. These guidelines, which are non-binding and are designed to evolve over time, explain the practical application of the legal concept, as anchored in the AI Act. The guidelines on AI systems definition were published in addition to the Guidelines on prohibited artificial intelligence (AI) practices, as defined by the AI Act. 

For the time being information on Commission’s activities in relation to article 4 of the AI Act can be found on the AI Pact webpage, including the recording of the webinar that took place on 20 February and the living repository on AI literacy. A dedicated webpage on AI literacy and skills is under preparation.

The AI Office will not impose strict requirements regarding Article 4 of the AI Act and its “sufficient level of AI literacy”. On the contrary, it considers necessary a certain degree of flexibility, considering the broad topic of AI literacy and the fast-evolving technology that AI is. Yet, as a minimum, to comply with Article 4 of the AI Act, providers and deployers of AI systems should:

a) Ensure a general understanding of AI within their organisation: What is AI? How does it work? What AI is used in our organisation? What are its opportunities and dangers?

b) Consider the role of their organisation (provider or deployer of AI systems): Is my organisation developing AI systems or just using AI systems developed by another organisation?

c) Consider the risk of the AI systems provided or deployed: What do employees need to know when dealing with such AI system? What are the risks they need to be aware of and do they need to be aware of mitigation?

d) Concretely build their AI literacy actions on the preceding analysis, considering

• differences in technical knowledge, experience, education and training of the staff and other persons – How much does the employees/person know about AI and the organisation’s systems they use? What else should they know?
• as well as the context the AI systems are to be used in and the persons on whom the AI systems are to be used – In which sector and for which purpose/service is the AI system being used?

Considerations a, b, c, and d include legal and ethical aspects. Therefore, connections to the EU AI regulation (i.e., understanding of the AI Act) and to principles of ethics and governance are encouraged.

As reported in the answer to the previous question, to comply with article 4 of the AI Act, organisations should consider their role (being providers or deployers of AI systems) as well as the risks associated to the AI systems they provide and/or deploy. According to this, organisations should adapt their AI literacy approach. For example, if the AI systems of the organisation are high-risk, according to Chapter III of the AI Act, additional measures might be relevant to ensure employees are aware of how to deal with the given AI systems and avoid and/or mitigate their risks.

This depends on the organisation’s answers to the considerations in question 1. Yet, in many cases, simply relying on the AI systems’ instructions for use or asking the staff to read them might be ineffective and insufficient. Article 4 of the AI Act is intended to provide trainings and guidance as most appropriate on the basis of each target group’s level and type of knowledge, as well as given the context and purpose of the AI systems in used in the organisation.

This is also in alignment with other provisions of the AI Act. For example, Article 26 introduces an obligation for deployers of high-risk systems to ensure that the staff dealing with the AI systems in practice is sufficiently trained to handle the system and ensure human oversight. Relying on the instructions of use is therefore not sufficient, further measures are necessary.

There is no one size fit all when it comes to AI literacy and the AI Office does not intend to impose strict requirements or mandatory trainings. The requirements for a training depend on the concrete context. While replicating the practices collected does not automatically grant presumption of compliance with Article 4, the initiatives in the living repository on AI literacy could provide some inspiration.

No, the AI Office does not impose requirements for specific sectors. Yet, as reported in the answer to question 1, the context – including the sector and the purpose – in which AI systems are provided/deployed should be relevant when developing and AI literacy initiative. Moreover, the level of risk of AI systems should be considered.

This depends on the concrete type of AI system and risk (e.g. for high risk). In general, people working for a service provider or contractor need to have the appropriate AI skills to fulfil the task in question (same as the employees).

Yes, they should be informed about the specific risks, for example hallucination.

These are two distinct questions. The employees and the human-in-the-loop need the appropriate skills, targeted as to the system they are using.

Normally yes, but it depends on the AI tool in question and their specific qualification. This is particularly relevant in view of the speed of the technological developments.

The organisation should still consider the steps in the answer to question 1 and asks itself: Do these technical employees know what needs to be known about the AI systems of the organisation and how to deal with them? Are they aware of all risks and how to avoid/mitigate them? Moreover, the organisation should consider what else these employees might need to know; e.g., legal and ethical aspects of AI.

Yes, article 4 of the AI Act encourages providers and deployers to consider the knowledge, experience, education and training of employees and other persons to provide a sufficient level of AI literacy. Given the difference between AI systems and that the level of knowledge and experience, as well as the type of education and training received, might vary, having different levels of training or learning approaches can be appropriate.

There is no need for a certificate. Organisations can keep an internal record of trainings and/or other guiding initiatives.

No, no specific governance structure is mandated to comply with article 4 of the AI Act.

The European Commission proposed the Digital Omnibus with targeted amendments to the AI Act, as part of the wider Commission’s effort to simplify and make more effective the EU’s digital rulebook. 

Regarding article 4, the Commission proposed to shift the obligation on the Member States and Commission to promote AI literacy and upskilling and reskilling in AI, rather than enforcing an unspecific obligation on organisations. By doing so, the Commission will also ensure continuity with the efforts so far undertaken (notably via this Q&A on AI literacy and the repository of AI literacy practices).

It is important to highlight that for those who deploy high-risk AI systems, the obligation to ensure that their staff is trained to ensure human oversight remains in place.

Article 4 of the AI Act entered into application on 2 February 2025, therefore the obligation to take measures to ensure AI literacy of their staff already applies. The supervision and enforcement rules apply from 3 August 2026 onwards.

The supervision and enforcement of Article 4 of the AI Act is not with the AI Office, but it is under the remit of national market surveillance authorities. The national market surveillance authorities will start supervising and enforcing the rules as of 2 August 2026. While decisions on enforcement are with the national market surveillance authorities in the different Member States, the AI Office will work closely with the AI Board to support an appropriate implementation.

The prohibitions apply since 2 February 2025. With the AI Board and cooperation between all relevant authorities, a coherent application of the rules will be ensured.

National market surveillance authorities could impose penalties and other enforcement measures to sanction infringements of Article 4. This will be based on national laws that Member States are due to adopt by 2 August 2025.

It is important to underline that the enforcement of the AI Act follows a proportionate approach. Any sanction must be proportionate, based on the individual case and take into account factors such as nature and gravity and the intentional and negligent character of the infringement. This might, however, be more likely if there is proof of an incident due to lack of appropriate training and guidance of employees or other persons.

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. This is also valid for Article 4 of the AI Act.

Public enforcement means that each Member State will nominate authorities responsible to enforce the provisions of the AI Act, who supervise and enforce compliance to this law. Private enforcement relates to the scenario where a private person suffers harm and considers that this is due to an organisation not complying with the obligation of Article 4 and sues the organisation to receive damages.

When there is a harm/damage, the provider of an AI system can be sued, according to national law. The AI Act doesn't create criminal offences or a right to compensation.

For the moment, guidance will be provided via further examples of practices, webinars and clarifications via this Q&A.

Further guidance on enforcement might be provided by the relevant national market surveillance authorities once nominated. The AI Office will work closely with the AI Board on the topic of AI literacy, in alignment with article 66(f) and article 95(2f) of the AI Act.

The Commission will publish guidelines on the application of the requirements and obligations referred to in Articles 8 to 15 and in Article 25 AI Act, and these guidelines will also touch upon issues of literacy, when discussing for example human-oversight or risk-management.

The AI@EC Communication already identified as operational action to Develop a policy to build and maintain an AI-skilled workforce. The European Commissions has already implemented several measures for its staff regarding AI literacy:

• The creation of internal AI specific web portal as one-stop shop accessible to all staff to the AI related content – AI guidelines, AI training resources, events, and news.
• Definition on the Commission training platform of AI learning packages, oriented to different targets – generalist, managers, and developers (specialist). These packages contain a curated list of relevant trainings, categorising then on essential, highly recommended and recommended.  Additional trainings and recording of webinars are available also in the platform.
• AI tools trainings - Specific section in the AI portal list the AI tools, available to all staff has been created that includes the relevant learning resources for each tool. There are periodic Q&A sessions on using AI in your daily work.
• An AI community of practice exists where any person can do questions related to AI and interact with AI experts.

Currently, many of the agencies have already access to the Commission’s learning platform (EU-Learn) as well as to resources such as the AI learning packages and other trainings of the Commission.

To support the implementation and compliance with article 4 of the AI Act, the AI Office will continue nurturing the living repository on AI literacy practices, gathering further examples from organisations, and updating the Q&A at hand. Further awareness activities will be organised and a dedicated webpage for activities related to AI literacy (within and beyond the remit of article 4), skills and talent will be launched with the aim of promoting access to AI literacy and foster dialogue on AI for all.

The AI Office highly values the insights and expertise of all stakeholders, including the industry. For this reason, we have created the AI Pact to foster the creation of a collaborative community, where stakeholders can share best practices and internal policies that may be of use to others in their compliance journey. With respect to AI literacy, within the AI Pact we have recently published a living repository of AI literacy practices; any provider and/or developer of AI systems that has put in place an AI literacy programme is invited to submit their contribution.

On a regular basis, the AI Office will verify that all contributions received meet the minimum criteria of transparency and reliability before accepting them into the public repository. During the verification period, the survey might be temporarily closed.

The living repository on AI literacy practices is freely available on our website: Living repository to foster learning and exchange on AI literacy

The repository includes several examples of ongoing initiatives to foster AI literacy organisations of different sizes providing and/or deploying different types of AI systems in various private and public sectors. The practices provide examples on how to approach the knowledge of different employees (e.g. technical – such as the one of developers, vs. non-technical, such as the one of an administrator or legal officer), but also refers to initiatives for other persons, including vendors, partners, clients. The list of practices is non-exhaustive and will be updated with further practices on a regular basis.

While replicating the practices collected in this living repository does not automatically grant presumption of compliance with Article 4, the repository aims to encourage learning and exchange among providers and deployers of AI systems.  By publishing the practices, the Commission does not imply either endorsement or evaluation in any way.

An example of EU initiative that could be of support to SMEs is the network of the European Digital Innovation Hubs (EDIHs). EDIHs are 251 one-stop shops across Europe (27 Member States, 3 EFTA countries – Norway, Iceland, and Liechtenstein; 8 associated countries: Albania, Bosnia and Herzegovina, Kosovo, Montenegro, North Macedonia, Serbia, Turkey, and Ukraine), supporting SMEs and public sector organisations to digitalise their processes. 80% of EDIHs already provide services focused on AI, including trainings, workshops, bootcamps for different needs and levels of knowledge of AI. SMEs and public sector organisations can contact EDIHs to receive support with AI technologies and their digitalisation goals more broadly. EDIHs can assist as first help desk on the AI Act and serve as entry points to other EU-funded supporting measures, such as, among others, Testing and Experimentation Facilities (TEFs) or AI Factories. You can find the EDIH closest to you on this catalogue.

In the context of the AI Pact, the AI Office hosts events and webinars to engage the AI community, promote knowledge sharing, and best practices to foster trustworthy AI. You can find all recordings of past events and the announcement of upcoming ones at AI Pact events.

An introductory training package on the fundamentals of the AI Act and the EU AI Innovation ecosystem’s instruments was recently made available on the EDIHs’ e-learning platform. While the package was conceived for EDIHs, it can be used also by a wider audience.

Further useful material on AI can be freely accessed via the Digital Skills and Jobs Platform, including: 

• Training offers  

• Learning Path  

• Learning Content

As there is no one-size-fits-all approach for AI literacy, there is also not one unique framework. Below are reported some of the most used frameworks at European and international level, according to their target group. 

For all citizens 

• The Digital Competence Framework for Citizen provides a common understanding of what digital competence is. The publication has two main parts: the integrated DigComp 2.2 framework provides more than 250 new examples of knowledge, skills and attitudes that help citizens engage confidently, critically, and safely with digital technologies, including AI. A new version of the framework will be released in 2025.

• The Digital Skills and Jobs Platform provides open access to a wide variety of high-quality information and resources for everyone interested in the broad topic of digital skills and jobs, including on the topic of Artificial Intelligence.

For the education sector, teachers and learners:

• These ethical guidelines on AI and data usage in teaching and learning are designed to help educators understand the potential that the applications of AI and data usage can have in education and to raise awareness of the possible risks so that they are able to engage positively, critically and ethically with AI systems and exploit their full potential.
• The European School Education Platform offers a number of materials on AI for teachers and learners.
• The community of practice of the Digital Education Hub offers an opportunity to education and training stakeholders to discuss and exchange on digital education topics, including AI. Within the Hub, a dedicated working group (squad) has published a report on AI in education providing further insights on AI skills development and teacher training.

• The UNESCO’s AI competency framework for teachers is intended to support the development of AI competencies among teachers to empower them to use these technological tools in their teaching practices in a safe, effective and ethical manner.
• The  UNESCO’s AI competency framework for students aims to serve as a guide for public education systems to build the competencies required of all students and citizens for the effective implementation of national AI strategies and the building of inclusive, just and sustainable futures in this new technological era.