Hotels and hospitality groups now use AI agents for guest messaging, reservations, pricing and staff rotas, and Article 4 of the EU AI Act applies to all of them. Operators must take measures so that staff and others acting for them have sufficient AI literacy. With seasonal hiring and agency staff, that means short, practical training before access, every time.
What does Article 4 ask of a hospitality business?
Article 4 applies to every provider and deployer of AI systems, high-risk or not. In substance, they must take measures to ensure, to their best extent, a sufficient level of AI literacy among their staff and other persons dealing with the operation and use of AI systems on their behalf, taking into account those people's knowledge, experience, education and training, the context of use, and the people the systems are used on.
AI literacy means the skills, knowledge and understanding to use AI systems in an informed way and to understand their opportunities, risks and possible harms. It does not demand that a receptionist understand model architecture. It does demand that they understand what the guest messaging agent may promise, and when to step in. Whether Article 4 requires formal training is covered in does the EU AI Act require AI literacy training.
Who does it cover in hospitality?
Hospitality structures make "who" a real question. A hotel may be owned by one company, operated by a management company and branded by a franchisor. The deployer is whoever uses the agent under its own authority, usually the operator. Its duty then covers:
- Front desk and reservations staff working alongside guest messaging and booking agents.
- Revenue managers setting the bounds of pricing agents.
- Housekeeping and food and beverage supervisors using agents that build rotas or allocate rooms to clean.
- Seasonal and agency staff who work under the operator's systems.
- Central reservations and outsourced contact centres acting on the operator's behalf.
- General managers and owners' representatives accountable for the property.
One use deserves care. Agents that allocate shifts or tasks based on individual workers' behaviour or performance fall within the Act's high-risk workforce uses, which brings Articles 14 and 26 for those overseeing them.
What does literacy mean when hospitality agents act?
Guest messaging and reservations. A governed guest agent might read bookings and the property's policies at depth 1, answer routine questions and confirm standard requests at depth 3, and have refunds outside policy, cancellations of non-refundable stays and compensation offers held for a named person. Front desk staff need to know which is which, and to recognise when the agent has promised something the property cannot deliver, such as an early check-in on a full night. They need to read where a fact came from: the property management system is sealed, a guest's statement is declared, and an email from a "travel agent" asking to redirect a refund is external. An external claim cannot authorise a payment.
Pricing. Revenue managers need to understand what drives the pricing agent, how to read the provenance of the demand data it uses, and how to pause it when an event or error makes its prices wrong. They also need to know where consumer pricing rules limit what it may do.
Rotas and task allocation. Supervisors need to understand what the agent optimises and what it cannot see, such as a staff member's caring duties or a room that needs extra work, and to override it. Where allocation is based on individual performance, that understanding becomes the competence Article 26 requires.
Guest data. Hotels hold passport details, payment data and preferences. Staff must know what the agent may use, and that a guest's data is not to be pasted into requests the agent does not need.
Lapses. An agent's certificate lapses after 90 days, and a run that clears no depth drops it to depth 1 the same day. On a busy weekend, guest messaging that suddenly needs a person for every reply is a staffing problem the duty manager must be ready for.
How does a hospitality business meet Article 4?
| Requirement | Control | Evidence it leaves |
|---|---|---|
| Sufficient literacy for staff | Short role modules for front desk, revenue, supervisors and managers | Records by role and property |
| Seasonal and agency staff | Induction module completed before system access | Induction records dated before first shift |
| Persons acting on the operator's behalf | Central reservations and outsourced centres trained under contract | Contract clauses; partner evidence |
| Awareness of risks and harms | Training on overpromising, refund fraud and guest data | Assessment results |
| Kept current | Briefings when agent policies or bounds change | Briefing records |
A step-by-step method is set out in how to set up an AI literacy programme under the EU AI Act.
What will a regulator ask?
Market surveillance authorities, data protection authorities and consumer authorities may ask: how are seasonal and agency staff trained before using agents? Who can override the guest agent's promises and refunds? How does the pricing agent's output get checked? How are rota decisions reviewed? Which company is the deployer at this property?
What does Article 4 not cover?
It does not set consumer, pricing or employment rules. It does not replace Articles 14 and 26 for high-risk workforce agents. It does not prescribe course content or credentials. It does not cover guests' own literacy, though Article 50 requires guests to be told when they are talking to an AI system. Governance questions for owners are discussed in the EU AI Act made AI literacy a board question; comparisons in Article 4 in consumer lending and in fintech.
This is an explanation of the rules, not legal advice; confirm how they apply to you with counsel.
Where the Academy fits
Pryme Intelligence Academy's lessons can be read or listened to as audio, with highlights and notes, which fits hospitality staff who learn between shifts. Lessons, labs and tests are free, and the Foundations badge is free with no expiry, which suits high-turnover teams. Our guide to AI literacy and certification for people who work with agents explains the levels.
Questions people ask
Is a franchisor responsible for franchisees' staff literacy?
Generally the deployer is the business using the agent under its own authority, often the operator. Franchise agreements should say who trains whom, and the operator should be able to evidence it.
How long should hospitality AI training be?
Short and specific. Front desk staff need to know what the guest agent may promise, what a held action means and who to call, which takes far less time than a general AI course.
Does a rota agent count as high-risk?
If it allocates shifts or tasks based on individual workers' behaviour or performance, it is likely within the high-risk workforce uses. Check the exact use with counsel.