Travel companies risk fines of up to 3% of their global turnover if they fall foul of new rules under the EU Artificial Intelligence Act. The provisions require clear labeling of AI-generated content and transparency when customers are interacting with an AI system, such as a chatbot.
The travel industry is already using AI-generated content widely. A recent study by AI forensic company ContentGuard.me and German digital marketing group ABCD Agency examined 25,550 hotel photos across seven destinations, finding that AI generation or heavy editing had been used in roughly one in five images.
Nick Parkinson, a partner at travel law firm Travlaw, said that computer-generated imagery and video is already an issue in the industry.
“Customers feel that hotel resorts are marketed using outdated images for resorts that have subsequently deteriorated. It could be the hotel itself, or the local area, such as the problem in Mexico with tons of seaweed piling up on beaches. If travel companies are to start using AI content to market resorts, it could amplify that risk,” he said.
Mathias Lundoe Nielsen, CEO of AI infrastructure platform Entravel Group, said, “I’d expect labeling requirements to hit budget and midmarket hotels harder than luxury ones. A five-star property already invests in professional photography because guests expect that level of polish regardless.
“It’s the more affordable hotels that use AI-generated images to look more competitive and attract bookings on tighter budgets, so they are more likely to have to disclose AI generation, and that disclosure could work against exactly the properties that need the extra edge to compete for guests.”
Lundoe Nielsen said that there are legitimate use cases for AI relating to imagery, for example, swapping a summer scene for a winter one to reduce seasonal marketing costs.
“The problem starts when it’s overused to the point where guests are booking based on a room or setting that doesn’t reflect reality, and that’s where it stops being a cost optimization and becomes unfair to the customer,” he said.
Industry preparedness
When asked about its readiness for the new regulations, the world’s largest travel company by market cap, Booking Holdings, said, “At the highest level, we are extremely supportive of the AI Act and the enhanced clarity it brings for companies.”
“Building trust with consumers is our top priority. And to that end, if you think about the product of travel, it's all about reassuring people that what they see and ultimately book online is going to be what they can expect in real life. That means that setting accurate expectations is incredibly important,” a company spokesperson told PhocusWire.
“While AI can bring new efficiencies, generating images and videos at scale (for example), that needs to be transparent. It also needs to be in service of the real-life experience.”
The EU has developed a series of icons that must be used when users are first exposed to content that has been partly or fully generated by AI. This includes text, images, video and audio.
This month also marks the start of enforcement action for other provisions in the EU AI Act, which was adopted in 2024. The act prohibits the use of AI systems in certain circumstances, such as offering differential pricing to different socioeconomic groups.
In 2024, Booking Holdings was fined €413.24 million by Spain’s competition authorities for alleged abuse of its dominant position on the Spanish markets. The company will be keen to avoid risking the wrath of European regulators on pricing in the wake of the AI Act coming into force.
The Booking spokesperson told PhocusWire, “We've been working on compliance with the AI Act for several years, so we have already been ensuring that how we leverage AI and machine learning in the consumer journey is transparent and helpful, whether that's in how we manage customer service, passing queries transparently from an AI voice assistant to a human being or how we help customers identify the right filters for their search via natural language processing.
“To be clear though, while it's our partners who set prices on Booking.com, we don't employ any price targeting tools based on socioeconomic data.”
Setting boundaries for AI use
Entravel Group’s Lundoe Nielsen said there is an important line between personalization and discrimination
“AI should help a traveler find the right option faster, not determine what they pay for the same room based on a profile of who they are, where they’re from or what a system thinks they can afford. Recommending is fine, but repricing based on profiling is not,” he said.
“I’d go further than most founders here: I think that kind of differential pricing based on personal data is close to an antitrust issue, not just an ethics one.”
The act also bans the use of AI for manipulative practices online.
“Nudge design that pressures someone into a booking decision they wouldn’t otherwise make is the same underlying problem as discriminatory pricing. It’s basically using what a system knows about someone against their interest rather than for it,” said Lundoe Nielsen.
The AI Act also specifies that companies cannot use AI to disadvantage candidates in any recruitment process, for example, by using AI to monitor emotions during an interview or to deselect potential employees for other discriminatory reasons.
Travlaw partner Nick Goodchild said, “There are documented cases where companies have trained an AI to weed through CVs based on existing employees, so then the AI ends up profiling stereotypes—e.g., it might assume all computer coders need to be white males from countries XYZ aged A to B. The bottom line is to make sure the AI system is well trained and tested before being used and, even after use, monitored to look for unforeseen AI decisions that could be discriminatory.”
The Booking spokesperson told PhocusWire, “Human involvement and oversight with AI is key, and that is particularly true with any use cases as part of the recruitment process. We do leverage AI in carefully monitored ways as part of our talent processes and tooling. In compliance with the law, this always comes with active human involvement and monitoring. It's an assistive tool, not a decision maker.”
Will companies really face huge fines for contravening the act?
“Those figures are for the biggest earners and biggest offenders,” said Travlaw’s Parkinson. “A small company inadvertently committing a minor offence should not expect to be put into liquidation. But a large company that is deliberately flouting some of the most important rules can expect a proportionate response.”