In Brief: The EU Artificial Intelligence Act introduces new obligations for travel companies using AI, including disclosure of AI-generated content, transparency around chatbots and limits on discriminatory pricing, manipulation and recruitment practices. Companies that breach major provisions could face fines of up to 3% of global turnover.
-
EU AI Act Raises Disclosure and Pricing Rules for Travel Companies – Image Credit Unsplash
AI-Generated Travel Marketing
The EU AI Act is expected to affect travel companies that use artificial intelligence for marketing, customer service, pricing, recruitment and search. Businesses must clearly label content partly or fully generated by AI and inform customers when they are interacting with an AI system, such as a chatbot or voice assistant.
The rules are significant for a sector that is increasingly using AI-generated images, videos and written content. A study by AI forensic company ContentGuard.me and German digital marketing group ABCD Agency reviewed 25,550 hotel photographs across seven destinations and found that about one in five images had been AI-generated or heavily edited.
Nick Parkinson, a partner at travel law firm Travlaw, said inaccurate visual marketing was already a concern for travelers. Hotels and resorts can be promoted using old images that no longer reflect the condition of a property or its surroundings, including seaweed accumulation on beaches in Mexico, he said.
Parkinson said AI-generated content could increase this risk if companies use it to market resorts in ways that do not reflect the actual experience available to customers.
Disclosure Requirements for Hotel Imagery
Mathias Lundoe Nielsen, chief executive of AI infrastructure platform Entravel Group, said disclosure requirements could have a greater effect on budget and midmarket hotels than on high-end properties. Luxury hotels generally invest in professional photography because guests expect detailed, polished images, he said.
More affordable hotels may be more likely to use AI-generated imagery to compete for bookings while managing marketing costs. However, disclosure that an image was generated by AI could affect how potential guests view the property.
Lundoe Nielsen said AI can have legitimate uses in hotel marketing, such as adapting an image from a summer setting to a winter setting without arranging separate seasonal photography. The issue arises when images depict rooms, facilities or surroundings that do not exist or do not match reality, he said.
The EU has developed icons that must be displayed when a user first encounters content that is fully or partially created by AI. The requirements cover text, images, audio and video.
Travel Companies Prepare for Enforcement
Booking Holdings supports the AI Act and the additional clarity it provides for businesses. A spokesperson said trust depends on ensuring customers see an accurate representation of what they will receive when they make a booking.
The company said AI-generated images and videos should be used transparently and remain connected to the real-world travel experience. Booking Holdings has been preparing for compliance with the law for several years.
The company uses AI and machine learning in parts of the consumer journey, including customer service and natural-language search tools. In customer service cases, users can be transferred from an AI voice assistant to a human representative.
The AI Act was adopted in 2024, and enforcement of some provisions is beginning. The regulation prohibits certain uses of AI, including systems that apply differential pricing to people based on socioeconomic factors.
Booking Holdings was fined 413.24 million euros by Spanish competition authorities in 2024 for allegedly abusing its dominant position in Spain. The company said accommodation partners set prices on Booking.com and that it does not use price-targeting tools based on socioeconomic data.
Limits on Pricing and Online Influence
Lundoe Nielsen said AI-based personalization should help travelers identify suitable options without changing the price of the same room based on personal profiles, origin or perceived ability to pay. He distinguished between making recommendations and repricing based on profiling.
The act also bans AI systems used for manipulative online practices. Lundoe Nielsen said digital design that pressures a person into making a booking they would not otherwise make can conflict with the same consumer-protection principles behind restrictions on discriminatory pricing.
Recruitment and Proportional Penalties
The AI Act also restricts the use of AI in recruitment where it could disadvantage job candidates. This includes emotion monitoring during interviews and selection practices that create discriminatory outcomes.
Travlaw partner Nick Goodchild said companies need to test and monitor recruitment systems for bias. AI trained using the profiles of existing employees could reproduce stereotypes involving gender, ethnicity, age or nationality, he said.
Booking Holdings said AI used in its recruitment processes is subject to active human involvement and monitoring. It described the technology as an assistive tool rather than a decision maker.
Parkinson said the maximum penalties are intended for serious breaches by large businesses. Smaller companies that make minor, unintentional errors are unlikely to face penalties that threaten their operations, while deliberate violations of important rules could result in proportionate enforcement, he said.
