AI is on a one-way ticket to upending the travel industry. 46% of American consumers expect it to improve their travel experience, and 49% of people who have tried AI say it’s “very effective” for planning a trip.
That’s according to a 2023 report from the National Research Group (NRG). They also cite journalists using AI to write travel guides and social media influencers who teach their followers how to plan vacations with it.
Major travel companies are adopting AI, too. Omri Morgenshtern, CEO of digital travel platform Agoda, recently predicted that generative artificial intelligence (GenAI) would, “be the biggest opportunity for business” in 2024. Airbnb plans to put AI front and center of its app, and Expedia recently launched its own AI trip-planning tool.
These developments raise worrying questions for a travel writer and blogger like myself. Since day one, AI has made headlines for disrupting industries. Will jobs like mine soon become irrelevant? Or is there still room for a human touch? I wanted to find out.
The most popular AI program, ChatGPT, clocks more than 1.5 billion visits every month. It’s also the same tool Expedia is using to help people plan their travels.
Users input questions or commands, and ChatGPT scrapes information already present on the web to deliver answers.
To assess the hype, I used my free ChatGPT account to plan a pretend two-week trip to London in July. I’d ask for recommendations on what to see and do, tips on how to structure my time, where to eat and sleep depending on my budget, and so on.
I’m from the U.K. and have lived in the capital, which I hoped would help me assess its suggestions. With luck, I’d gain a better sense of how useful AI currently is for trip planning.
In my initial prompt, I wrote, “I’m going to London for 2 weeks. Give me a list of 10 essential things to do while I’m there.” It probably took three seconds to provide them. I decided 10 wasn’t enough, so I typed, “Anything else?” and hit enter.
The first noticeable perk of using ChatGPT over a traditional internet search was speed. Just like that, I had 20 potential things to do, including day trips from London. The list didn’t include a few major attractions, such as Westminster Abbey and St. Paul’s Cathedral, and the ideas were unoriginal — “Experience the West End” and “Tour the Tower of London.” Yet they were still well-known and worthwhile activities.
Moreover, the idea-generating process was frictionless. Users submitting queries don’t have to click anything or scroll through distracting ads. The answers came instantaneously.
ChatGPT’s next trick was even more impressive. I asked it to “create an itinerary for my 2 weeks there, including how long it takes to get between attractions and how to get around.”
In record time, it broke my trip into four parts with distinct themes — the first three days were “Central London Exploration” — and offered plans to fill each morning, afternoon, and evening. The suggestions made sense directionally, and the timing seemed to track.
At first glance, I couldn’t fault the suggestions. However, closer inspection revealed mistakes; ChatGPT had suggested activities that wouldn’t have worked.
For example, the service’s afternoon activity recommendation for Day 3 was to “Visit Buckingham Palace and watch the Changing of the Guard ceremony.” According to Visit London, the city’s official visitor guide, this only happens at 10:45 a.m. on Mondays, Wednesdays, Fridays and Sundays. Similarly, the morning activity slated for Day 6 was to “Explore the Columbia Road Flower Market,” which runs exclusively on Sundays.
OpenAI, the creator of ChatGPT, admits the tool can make mistakes and encourages users to verify information.
ChatGPT will presumably make fewer mistakes as it improves over time, but can today’s users trust its responses if a third party must verify its suggestions? Why use the service in any capacity if site visitors utilize travel articles for fact confirmation?
The prompts matter, too. For example, I asked, “Do you have any recommendations for where to stay? I’m traveling on a budget.”
I wanted the names of affordable hotels in London. Instead, ChatGPT gathered a list of cheap accommodations, like “hostels” and “Airbnb.” I had to narrow my search by typing, “Any specific hotels you recommend?” to get the desired outcome.
Credit where it’s due, though, as the right prompts can deliver impressive results. For instance, I tried a similar prompt to one I’d seen in a viral video on this topic:
“I will be visiting central London in the U.K. for two weeks in July. Create me a table that breaks down top tourist activities, destinations, and dining experiences. The columns should be name, type of attractions, cost for two people, description of activity, and distance in miles from central London. Have all options be easily accessible by public transport.”
Moments later, I was looking at the graph. Some details weren’t perfect, and there were a few unusual suggestions, such as visiting a train station. Still, it took seconds to create a well-structured table packed with helpful information — a task that could have taken hours.
My ChatGPT experiment delivered mixed results. In many ways, the tool was extraordinary. In others, it was reassuringly average.
As a travel writer, I found the experiment both exciting and demoralizing. On the one hand, ChatGPT could inform and expedite the writing process. On the other hand, it makes human writers less necessary, especially for topics requiring minimal expertise or firsthand knowledge.
My main reason for hope comes from my love of travel. I can imagine ChatGPT becoming a go-to tool for vacationers with time constraints seeking quick trip ideas to accommodate last-minute plans.
Yet, trip planning isn’t always a chore. For people like me, it can be part of the fun. I don’t want to reduce the process to text and tables. I want to research the destination, find its hidden gems, and see images that help me imagine myself there. ChatGPT takes all that away.
Ultimately, the clue is in the name. AI is exactly that: artificial. It’s like the microwave meal of travel planning – lightning-fast and convenient but lacking substance.
For now, I expect many consumers, particularly travel enthusiasts, to continue seeking suggestions from real people they like and trust and who can make recommendations based on recent, firsthand experiences. Time will tell if I’m right.
Boutique Travel VC raises investment cap from $1.5M to $10M to lead rounds and support portfolio companies through growth Travel...
Read moreDetails
