AI literacy under Article 4 of the AI Act: what teams running robots must do

Article 4 of the EU AI Act requires providers and deployers of AI systems to take measures that support the AI literacy of their staff and of other people who operate or use those systems on their behalf. It has applied since 2 February 2025, national authorities have supervised it since 2 August 2026, and the AI Omnibus reworded it in July 2026. It still applies to robot teams.

This explainer sets out what the article says today, what the Omnibus changed, who it covers, and how it applies to humanoid robots, robot dogs and service robots at events. It is general information, not legal advice.

What Article 4 says today

The consolidated text published on the Commission's AI Act Service Desk page for Article 4 has three paragraphs, all marked as amended.

• Paragraph 1 asks providers and deployers to take measures to support the development of AI literacy among their staff and others dealing with the operation and use of AI systems on their behalf. They should take into account those people's technical knowledge, experience, education and training, the context in which the systems are used, and the people the systems are used on. The text now states explicitly that this does not oblige anyone to guarantee a particular level of AI literacy for any individual.
• Paragraph 2 asks the Commission and Member States to support providers and deployers in meeting the obligation, with particular attention to SMEs, and has the Commission publish practical examples on a single information platform.
• Paragraph 3 asks the European AI Board to adopt recommendations on promoting AI literacy, taking existing competence frameworks into account.

The AI Act also defines the term. According to Article 3 on the Service Desk, point 56 describes AI literacy as the skills, knowledge and understanding that let providers, deployers and affected persons make informed use of AI systems and be aware of AI's opportunities, risks and possible harms.

What the AI Omnibus changed

The Commission proposed the AI Omnibus as part of its digital simplification package on 19 November 2025. According to the Commission's announcement AI Omnibus enters into force, dated 27 July 2026, the amending regulation took effect on that day. The Commission says the AI literacy rule was simplified so that the Commission and Member States take a larger role in promoting AI literacy, instead of companies carrying the burden alone.

This change is easy to misread. The Commission's AI Literacy Questions and Answers, last updated on 20 October 2026, says AI literacy remains an obligation for providers and deployers. What has gone is any requirement to reach a set or sufficient level, which was the wording of the original text. The same page says that deployers of high-risk AI systems must still make sure their staff are trained to carry out human oversight.

Put simply, the Omnibus made the duty less specific about the result. It did not remove it.

Who is covered

Article 4 applies to two roles that the AI Act defines in Article 3:

• A provider develops an AI system, or has one developed, and places it on the market or puts it into service under its own name or trademark, whether it charges for it or not.
• A deployer is any person or organisation using an AI system under its authority. The only exception is use in a personal, non-professional activity.

The Commission's Q&A adds three points that matter in practice:

• The duty covers all AI systems, not only high-risk ones, but the effort should be proportionate to the risk.
• People acting on behalf of the organisation are not only employees. The Commission gives contractors, service providers and clients as examples.
• Organisations outside the EU are covered when the AI system is placed on the EU market or used in the EU, or when its use affects people in the EU.

Does a robot count as an AI system?

Under Article 3(1), an AI system is a machine-based system that operates with some degree of autonomy, may adapt after deployment, and infers from its inputs how to produce outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments.

Many robots now used at events contain components that plausibly fit this description: speech recognition, conversational assistants built on language models, camera-based perception and autonomous navigation. Other functions may not. A pre-programmed dance played back step by step, or a robot steered entirely by a remote operator, may involve little or no inference of this kind. Our guide to teleoperated and autonomous robots explains how these modes differ on the day.

The answer therefore depends on the software actually running during your event, not on whether the robot looks humanoid. Whether a robot's AI also falls into a high-risk category is a separate question, covered in our EU AI Act timeline for robots.

Who is the deployer at a robot event?

The manufacturer that develops the robot's AI and sells it under its own brand will usually be the provider. When the robot is hired, it is less clear who the deployer is. The answer depends on who uses the system under its own authority.

• A rental supplier that brings the robot and runs it with its own operators is likely to be using the system under its own authority, so it would need to cover its operators.
• A brand or event organiser that writes the robot's conversational content, configures its behaviour, or has its own staff run it on a stand may also be using the system under its authority.
• Agency staff, freelancers and venue crew working with the robot on behalf of either party can fall within the circle of people whose AI literacy should be supported, since the Commission names contractors and service providers.

It is worth agreeing in writing who configures and controls the AI functions. Our guides on who operates the robot and what staff you need and what to put in a request to a robot supplier cover the practical side.

What supporting AI literacy looks like

The Commission's Q&A says there is no one-size-fits-all approach, no mandatory training and no certificate. It suggests four steps:

1. Build a general understanding of AI within the organisation.
2. Establish whether the organisation is a provider or a deployer of the systems concerned.
3. Consider the risks those systems carry.
4. Design literacy measures around people's existing knowledge and the context in which the systems are used.

Two further points from the Q&A are relevant. First, it warns that in many cases simply relying on the instructions for use, or asking staff to read them, may not work. Second, it suggests keeping an internal record of training and other guidance as evidence. For staff who use general-purpose chat tools, it names the risk of hallucination, meaning confident but incorrect output, as something they should understand.

For a robot deployment, measures in this spirit could include briefing operators and stand staff on:

• what the robot's speech system can get wrong, such as invented answers or misheard questions, and how to correct it in front of visitors;
• which functions are autonomous and which are controlled by a person, so that no one overstates what the machine can do;
• the limits of obstacle detection and when to use the stop function, in line with our guide to robot safety at public events;
• what the cameras and microphones record and where that data goes.

These are examples, not a checklist set by the regulation.

Enforcement and penalties

According to the Commission's Q&A, national market surveillance authorities have supervised and enforced the rules since 2 August 2026. The Q&A also says that any penalties should be proportionate and decided case by case, and that enforcement is more likely where an incident can be traced to a lack of training and guidance.

The AI Act does not set a specific fine for Article 4. The fine levels in Article 99 name other provisions, and paragraph 1 leaves it to Member States to set the rules on penalties and other enforcement measures for breaches of the Regulation. National approaches may therefore differ.

For high-risk systems, a stricter rule applies alongside Article 4. Article 26 requires deployers to give human oversight to people with the necessary competence, training and authority. According to Article 113, as amended, most high-risk obligations apply from 2 December 2027 or 2 August 2028, depending on the category.

What this means for event organisers

• Treat Article 4 as a current obligation. It has applied since February 2025 and is now supervised. The Omnibus removed the target level but kept the duty.
• Establish your role. If your team configures or runs the robot's AI functions, you may be a deployer in your own right, not only the supplier.
• Ask suppliers how they prepare operators. Find out what operators know about the robot's AI functions, their limits and the stop procedures, and whether training is recorded.
• Brief your own people. Hosts, promoters and agency staff working beside the robot should understand what it can and cannot do, so that they neither overstate it to visitors nor rely on it in unsafe ways.
• Keep a short record. A dated note of who was briefed on what is a simple way to show the effort the Commission describes.

Frequently asked questions

When did the AI literacy obligation in Article 4 start to apply?

Article 4 sits in Chapter I of the AI Act, which has applied since 2 February 2025 under Article 113. According to the European Commission's AI literacy Q&A, national market surveillance authorities have been responsible for supervising and enforcing it since 2 August 2026. The AI Omnibus, in force since 27 July 2026, changed its wording but kept the obligation for providers and deployers.

Did the AI Omnibus remove the AI literacy obligation for companies?

No. According to the Commission's Q&A, AI literacy remains an obligation for providers and deployers of AI systems. The amended text no longer sets a specific or sufficient level, and it gives the Commission and Member States a stronger role in promoting AI literacy. Deployers of high-risk AI systems must still make sure the staff who carry out human oversight are properly trained, under Article 26.

Do staff need a certificate to show AI literacy under the AI Act?

No. The Commission's Q&A says no certificate is needed and no particular training is mandatory. It suggests that organisations keep an internal record of training sessions and other guidance they have provided. The measures should be proportionate to the risks of the AI systems involved and to the knowledge and role of the people using them. Simply handing out the instructions for use may not be enough.

Does Article 4 apply to contractors and agency staff working with a robot?

It can. Article 4 covers staff and other people who deal with the operation and use of AI systems on behalf of a provider or deployer. The Commission's Q&A gives contractors, service providers and clients as examples. At an event, this could include freelance operators, agency hosts or venue crew working with the robot's AI functions. Who is responsible depends on who uses the system under their own authority.

How this article was made: written by the humanoid4rent editorial team with AI assistance; every factual statement was checked against the sources listed below on 2026-10-31. Spotted an error? Write to info@humanoid4rent.com.