For many travelers, the process of planning a trip is a daunting task of endless price comparisons and cluttered browser tabs. More and more tasks are being outsourced to artificial intelligence.
According to an annual survey conducted by travel platform Klook of 11,000 users around the world, around 91% of global travelers use AI travel planners.
Travelers use this technology for a variety of reasons. According to Klook’s research, some people rely on AI to tell them what they want from a trip, while others use it to find the best deal for their needs.
Lee Rowan, from travel agency Savanti Travel, said the appeal partly reflects “a do-it-yourself culture where you feel a sense of ownership and pride in creating something yourself.”
But trust remains a challenge. Another Booking.com report on consumer attitudes towards AI in daily life and travel found that 91% of respondents remain concerned about AI, and only 35% fully trust its output.
Over-the-shoulder view of young Asian woman using smartphone to check digital airline ticket and boarding pass on device screen while packing suitcase on bed for travel. Planning a trip. travel and vacation concept
D3 Sign | Moment | Getty Images
“Illusion” hurdle
One of the major hurdles is accuracy. AI tools built around large language models are known to cause “hallucinations,” a phenomenon in which false information is presented as fact.
Shin Yi Ho, director of tourism consultancy Horwath HTL, said her AI experience had few hallucinations.
When Ho, a self-described “heavy user” of LLMs, was planning a six-month solo sabbatical, she relied on ChatGPT’s suggestions to find hotels in the destination that fit her interests and within her budget. Those recommendations were “very clear and good,” she said.
Although Ho’s experience with AI has been positive, the reliability of AI remains a concern for many travelers.
Savanti Travel’s Rowan recalled that a client who was in Paris on business arrived late at the appointed time because the route suggested by ChatGPT did not take into account road closures due to construction. What was supposed to be a 10 minute trip turned into a 45 minute trip.
“They seem like special cases, but they’re actually very common,” Rowan said.
AI paradox
Industry experts say AI could also change the destinations travelers see. Small and medium-sized businesses can be particularly affected, as they lack the expertise and digital presence needed to appear in search results.
Horwath HTL’s Mr Ho said older properties, particularly independent properties, and even properties in developing countries, “will struggle even more because there is no expertise or guidance”.
“That’s unfortunate,” she said. “Because they definitely need the demand more than ever, right?”
Conversely, well-promoted tourist destinations may suffer from the effects of overtourism with current versions of AI.
Guy Llewellyn, assistant professor at EHL Hospitality Business School Singapore, said AI tools should be used to introduce off-beat places and spread demand to avoid the risk of overtourism. Many AI systems are trained on a “top 10 list” of places to recommend, which is a bit contradictory, he added.
Crowds of tourists gather on the street near Kiyomizu-dera Temple in Kyoto, April 2019.
Nicola Datish | Light Rocket | Getty Images
AI tools can also struggle with real-world nuances that experienced travel planners consider. Rowan said this includes considering seasonal weather when proposing outdoor experiences, as well as considering travel fatigue after a long flight.
“Multi-generational mobility issues, obvious allergies, disabilities, intolerances are things that AI doesn’t really manage very well,” he said.
It also can’t help you if things go wrong, he added. When war broke out in the Middle East, airspace in the region was closed and many travelers were stranded. AI won’t get you to the front of the line for the next repatriation flight, he says.
The road ahead
But as AI models improve and companies become more adept at integrating new tools, Llewellyn is hopeful about their usefulness.
“As they become more integrated with AI, they may actually have access to more up-to-date data than traditional travel agents,” he added.
Llewellyn said the hospitality industry needs to prioritize structuring and publishing data so that it can be accessed by AI.
“It doesn’t have to be customer-facing; it can be purely on the back end through API calls,” he added, referring to “application programming interfaces,” or gateways that allow outsiders to use a company’s data for various purposes.
To the extent that AI is accessible, it could provide factual information to end users and reduce overall hallucinations, he said.
Despite information gaps and flaws in existing AI models, the hospitality industry is rolling out smarter tools. For example, Booking.com uses OpenAI to implement a variety of AI-powered solutions.
Llewellyn predicts that the digitization of information will continue to proliferate as more AI companies enter the market.
Travelers should benefit from this by reducing the frequency of hallucinations and improving the overall user experience.
“AI planners will emerge,” Llewellyn said. “The first few iterations will take time. There will be some hiccups, but it will have a huge impact on the industry.”
