Training requirements · EU AI Act

Is AI literacy training mandatory under the EU AI Act?

Yes, in effect. Every organisation that provides or uses AI systems in its work must take measures to support the AI literacy of its staff and of others who operate those systems for it. The law does not prescribe a course, an exam or a certificate, but supervisors expect to see what you did, so structured training with records is the usual way to show it.

Key facts

Applies to
Providers and deployers of AI systems, including organisations outside the EU in some cases
Who must be covered
Staff and other persons operating or using AI systems on the organisation’s behalf
In application since
2 February 2025; wording amended with effect from 27 July 2026
Supervised by
National market surveillance authorities, from 2 August 2026
Certificate required
No
Training format required
None prescribed; measures suited to the people and the use

Checked against the official text on 16 November 2026

Who does the AI literacy duty apply to?

The duty applies to two kinds of organisation under the EU AI Act: providers, which develop an AI system or have one developed and place it on the market or put it into service under their own name, and deployers, which use an AI system under their authority in their professional activity. In practice that second group covers almost every organisation that has given its staff an AI tool for work: a chatbot in customer service, a writing assistant in marketing, a screening tool in recruitment, an AI feature inside a business application.

The Act can also reach organisations outside the EU, for example where they place AI systems on the EU market or where the output of their AI system is used in the EU. A provider or deployer in the Gulf, India or South-East Asia that serves EU customers should therefore check whether the duty applies to it, and most will find it simpler to meet it than to argue that it does not.

The people covered are the organisation’s own staff and other persons dealing with the operation and use of AI systems on its behalf, such as contractors and service providers. The European Commission’s guidance notes that, depending on the risk, it can also be useful to build literacy among clients who use the system.

What changed in July 2026?

The duty has applied since 2 February 2025. As first adopted, it required providers and deployers to take measures to ensure, as far as they could, a sufficient level of AI literacy among the people concerned. The Digital Omnibus on AI, adopted on 8 July 2026 and in force from 27 July 2026, rewrote it. Providers and deployers must now take measures to support the development of AI literacy, and the text states that this does not require them to guarantee any specific level of AI literacy of any individual. The Commission and the Member States must support these efforts, in particular those of small and medium-sized enterprises.

The change lowers the bar from a result to an effort, but it does not remove the duty, and supervision began on 2 August 2026, when national market surveillance authorities started to supervise and enforce it. An organisation that has taken no measures at all is in the same position after the amendment as before it.

What must AI literacy training cover?

The Act defines AI literacy as the skills, knowledge and understanding that allow people to make an informed deployment of AI systems and to be aware of the opportunities and risks of AI and the harm it can cause. The measures must take into account the people’s technical knowledge, experience, education and training, the context in which the AI systems are used, and the persons or groups the systems are used on.

That means one generic video for everyone is rarely enough. A useful programme has a common base for all staff and role-specific depth on top of it:

  • What AI is and is not, and how the AI systems your organisation actually uses work, at a level each role can follow
  • What each system may and may not be used for in your organisation, and who decides
  • How to read and check an AI output: errors, made-up answers, bias, and when a human must review the result
  • What data may be entered into an AI tool, and what must never be, including personal and confidential data
  • The risks to the people the system is used on, and how to raise a concern or report an incident
  • For managers and owners of AI systems: the organisation’s AI policy, risk assessment and the obligations that apply to its uses of AI

What evidence will a supervisor or auditor ask for?

There is no prescribed format and no certificate requirement. The Commission’s guidance says that organisations can keep an internal record of the training and other guidance they provide. In practice, a market surveillance authority, a client’s due-diligence questionnaire or an ISO/IEC 42001 certification auditor will look for the same things:

  • An inventory of the AI systems in use, and of the roles that use or oversee each one
  • A short AI literacy plan: which roles get which training, and why it fits their use of AI
  • Training records per person: course, date and result
  • The AI use policy or rules staff were trained on, with its version and date
  • Refresher training when systems, uses or rules change
  • For high-risk AI systems, evidence that the people assigned to human oversight have the competence, training and authority the Act requires of deployers

How do you build an AI literacy programme that stands up to scrutiny?

Most organisations can put a defensible programme in place in a few weeks. The steps below are the ones Sterling’s experts use with clients building AI governance, whether or not they plan ISO/IEC 42001 certification.

  1. List the AI in useRecord every AI system and AI feature staff use, including tools bought as a service, and who uses or oversees each one.
  2. Set the rulesWrite or update a short AI use policy: approved tools, permitted uses, data that must never be entered, and when a human must review the output.
  3. Match training to rolesGive every AI user a common base course, then add depth for managers, system owners and anyone assigned to oversee a high-risk system.
  4. Train and recordRun the training, and keep per-person records of the course, the date and the result in one place.
  5. Review and refreshRevisit the plan when systems, uses or rules change, and at least at each management review, and retrain where the change matters.

What happens if an organisation does nothing?

The AI literacy duty does not carry its own fixed EU fine level; penalties for breaches are set by each Member State in national law. The larger exposure is indirect. An organisation that cannot show any AI literacy measures will struggle to answer a market surveillance authority, a client’s supplier questionnaire or a certification auditor, and it carries the business risk that the duty exists to reduce: staff entering confidential data into public tools, decisions based on unchecked AI output, and harm to the people the systems are used on.

How does Sterling Academy help you meet the duty?

Sterling Academy’s AI courses are built by industry experts, domain specialist trainers and auditors with 30+ years in ISO and GRC, in the same content and format as Sterling’s classroom training. They follow the current text of the EU AI Act as amended in July 2026. Every learner ends with a final exam and a verifiable certificate, and the client administrator sees each person’s progress and result in one place, which gives you the training record a supervisor or auditor asks for.

A typical programme combines an AI literacy course for all staff who use AI tools, an EU AI Act awareness course for managers and the people who own AI systems, and ISO/IEC 42001 training for those building an AI management system. The courses below are live and can be bought for one person or for a whole team.

Frequently asked questions

Questions about EU AI Act training

Ask a Sterling consultant

Is AI literacy training a legal requirement in the EU?

Providers and deployers of AI systems must take measures to support the AI literacy of their staff and others operating AI on their behalf. Training is not named as the only measure, but it is the most common way to meet the duty and to show that you did.

Do employees need an AI literacy certificate?

No. The EU AI Act does not require a certificate. A certificate from a final exam is still useful evidence for your own records, for clients and for auditors.

Does the duty apply to companies outside the EU?

It can. The EU AI Act reaches some providers and deployers outside the EU, for example where they place AI systems on the EU market or where the output of their systems is used in the EU. Check your own position with your legal adviser.

What changed with the Digital Omnibus on AI?

From 27 July 2026 the duty reads as an obligation to take measures to support the development of AI literacy, and it no longer requires a guaranteed level for any individual. The duty itself, and the people it covers, stayed.

Who enforces the AI literacy duty?

National market surveillance authorities in each Member State, which began supervising and enforcing it from 2 August 2026. Penalties are set in national law.

How often should AI literacy training be repeated?

The Act sets no fixed interval. Repeat or update the training when you introduce new AI systems, change how they are used or change your AI rules, and keep a record each time.

Is AI literacy the same as the training required for high-risk AI?

No. It is the base for everyone who uses AI. Deployers of high-risk AI systems must also assign human oversight to people with the necessary competence, training and authority, which needs deeper, system-specific training.

Sources

Prepared by the Sterling expert panel · Last updated 16 November 2026 · How our courses are made · This page is general information, not legal advice.

Training your team?

Plan the training
with a Sterling consultant.

Discuss your training needs