We have seen a lot of articles about AI and travel recently, from companies trying to push it onto us against our wishes to tales of it going horribly wrong. And of course, you can’t go anywhere without people mentioning how it will be the great event of our age. So, we thought we’d spend a little bit of time explaining how AI is affecting the travel industry and why, very truthfully, we are not concerned. Ignoring for now the environmental impact of AI and the ethics of illegal data scraping, this article is just about AI and travel, and how the hard limits of the technology become apparent whenever you aim higher than ‘fine’.

‍

But first, we want to clear the air. AI, as an emerging technology, has clear benefits that can be used. Like the computer or the internet before it, each advance in technology brings along an opportunity to improve the way we do things. At The Explorer Society, AI assists us with certain aspects of our business. We built our entire back-end system from scratch, and when AI came along we looked at ways to implement the technology to make our service better.

‍

More important is what we don’t use it for. Our itineraries are crafted by a person, with the assistance of other people around the world. This article and all written information you read from us (like guides or social media) are written by a person, and edited by another person. The difference between the work we get AI to assist with, and the work we do ourselves? Trust and creativity.

‍

On Trust and Being 'Confidently Wrong'

Trust is important, as our travellers come to us to help design their dreams. They need to know that their trip has been designed based upon accurate, up-to-date information, is well suited to their needs and has been made by a real person with real lived first-hand experience.

‍

AI is great at giving the illusion of authority. But for anyone that has used AI more than a few times, it quickly becomes apparent that AI chat bots are most concerned with giving you a quick answer that best confirms the question you've asked, not the answer that is most accurate.

‍

There was a recent warning on Smartraveller, the Australian government travel advisory, about using AI for planning travel. Whilst a useful tool for research, it warned that it can’t be trusted to provide accurate information to plan a trip around. AI usage has presented countless examples of the technology being ‘confidently wrong’.  Like motorists blindly following Google Maps directions into a river, sometimes AI will come up with information that is not at all correct. It can be handy to have the answer quickly, but that answer is worthless if you can’t trust it. We, out of curiosity, tested AI tools on visa requirements and local country information, and on many occasions it gave obviously incorrect information. This included information that was no longer applicable, or just a fundamental misunderstanding of geography. For something as clear and important as this, there is no excuse for trading trust for speed, especially when we have decades of lived experience and people on the ground around the world to ask instead.

‍

A great example of this happened to us only recently. A family member of ours recently visited Australia with an itinerary she had partially planned out with ChatGPT. She had told it that she wanted to see koalas and kangaroos and to visit some major sights. The trip that the AI spat out needed some serious correction. It involved bouncing all around the country (often impossibly) and inevitably spending more time in cars than on location. To see kangaroos, it recommended driving a few hours to a football oval in rural Victoria. Whether people had seen kangaroos there once prior or the local team was called the Kangaroos, we weren’t sure. But, for someone travelling internationally for a holiday, that’s a fundamental error.

‍

On Creativity and 'AI Slop'

Creativity in the design of any itinerary is just as important, and is often the overlooked element here. AI models work by taking a colossal library of digitally scraped data to train their systems, or language models. As is quickly evident whenever you take a huge sample size of anything, the results usually end up being very general, or middle of the road. If you were to make a soup using ingredients that 10,000 other people put in their soup then you are going to end up with something hot and wet, but not great tasting. The same applies to great travel itineraries. If you are using AI to write a report around quarterly earnings, then no-one is looking for high art. But, if you are planning out a once-in-a-lifetime trip for real people with real dreams and real wishes, then middle of the road isn’t good enough.

‍

This is quite clear to anyone spending a few moments on AI. Some travel magazines recently ran tests using AI to plan trips for them, and the most common result was ‘it was fine’. It filled the brief, but wasn’t anything special. It was sampled from the thousands of online travel aggregators, and countless Reddit posts, and random TripAdvisor forums, and what came out was usually really generic in nature. It is the sort of stuff you see on cheap package tours, where actually experiencing the destination is less important than simply ticking it off your bucket list. What these itineraries never end up being is personal, or tailored, or unique. If you are looking for a 5-night beach getaway in Bali then by all means go generic, but if you are planning a dream trip through Africa then you might want to aim higher. This mass-market itinerary creation is the travel version of “AI slop”.

‍

The Hype Versus the Reality

We think AI makes a great assistant, for some tasks. We treat it like an enthusiastic student, full of studied knowledge but with no real-world application. Rather than being a fundamental reshaping of an industry, it is merely another tool in the toolbox. It sits in there next to email, and spreadsheets, and PDFs. Hype doesn’t equal reality, and never forget that all of those who shout about AI being the next best thing usually have some money in the game. After all, how many of us use crypto to buy new NFTs before taking a Segway home to spend time in the Metaverse? Weren’t all of these things undeniably the future, and we all had to get onboard or be left behind?

‍

And now even the loudest proponents of AI are starting to quieten, as the general public realise that AI can be a useful tool for facilitation but isn’t going to replace real-world expertise, or creativity. There is skill and imagination in all great things, from art to music, and AI can only replicate but not create. If you want to generate a quick throwaway joke image for your friend then AI works great, but you’d never hang one in the Louvre. That fundamental difference between Generated and Crafted is the great divide that AI can’t bridge. Anything of great value is made with love and care, be it a handwritten card or a work of art. Real travel, the type we love, is no different. These life-changing trips are designed for you by living experts, with immense care.

‍

The result is a different kind of AI: one of Amazing Itineraries, not Artificial Intelligence.