Recently, I had the opportunity to participate as a guest expert on Latvian Radio 1’s show “Kā labāk dzīvot” (How to Live Better), where we discussed the application of Artificial Intelligence (AI) in travel planning. The full recording of the show is available here. I’ve been familiar with AI tools for some time, even before the generative AI boom, as I’ve been working with them in the MedTech sphere since 2018, and more recently with typical office tools like Copilot. This time the application is different, but the basic principles are the same. There are things that can be done and things that cannot. In travel planning, I recommend viewing such a tool just like any other information available online, because it can also be inaccurate:

• People in forums also tend to give incorrect advice: “Feel free to bring fruit from Thailand, no one checked anything for me” (even though that is completely wrong).
• Google Maps can also have old info about opening hours, just like any website (I once corrected the opening hours of El Fishawy cafe in Cairo—it’s been open since the time of Napoleon, so it was unlikely to have closed exactly during our visit—and it hadn’t, it was most likely a competitor who had meddled).
• Blogs and press articles also tend to have wrong, inaccurate, and outdated information—prices change, laws change, and finally, there are people who write about places they haven’t even visited.

On the show today, I shared a couple of my own examples of how successful I’ve been using AI tools:

• The relatively positive: For a trip to New York, I entered a prompt stating that it would be my seventh time in the city, that I had seen all the main sights, that the weather would be bad, and what type of objects I was interested in. I received a suggestion to visit The Morgan Library, which I hadn’t seen on any other recommendation list before. Unfortunately, I also received a bunch of suggestions for various museums and sights that no longer exist. I also received an afternoon tea suggestion for the exact place I had already chosen myself.
• The not-so-positive: For a trip to Mauritius, I was looking for sights while at the airport because we found out about the trip shortly before departure. At first, the suggestions seemed great—the sights were indeed ones I wanted to see. However, upon closer inspection, it turned out that the daily plans weren’t very logical—they didn’t match the actual working hours and involved driving too great a distance in one day, driving around the whole island almost every day instead of focusing on one region per day. Also, when exporting the data in table format, the entries suddenly disappeared from the table, even though they were just in the list. Therefore, I used it as a quick summary of what’s worth seeing, but not for the final plan.
• The negative: A while ago, I was looking for case law precedents for a complaint I was writing about delayed luggage. Unfortunately, the AI offered completely non-existent cases as evidence, as they were not mentioned in any of the references it provided. When I pointed this out, I received a response that it had made a mistake. That happens—complete hallucinations. It’s no wonder a new term has also appeared—slop, which describes low-quality AI products. And it’s no wonder that, for example, consultants have now appeared in software development to help fix what was generated by AI for those who do it without knowledge.

When I tried to find tickets at a more favorable price, the results were unconvincing—believable for some destinations, but about 6 times higher than what could actually be found for others. I currently find tickets better myself. However, for things like translation while traveling—it’s a super solution! And that is also AI. Therefore, for those who say they don’t use it at all—that’s foolish. It’s time to start learning to use these tools, because I believe there will be spheres (for example, loyalty points) where they will soon be indispensable.

It is worth remembering that models are being improved, and quality is getting better. There isn’t one single best one, because their strengths differ. If you use the free version, Gemini also uses current data, unlike the free ChatGPT, but you still can’t take everything as 100% truth. Just like anything else published. Because in articles, too, perhaps the top 10 is mentioned precisely because the author receives affiliate marketing commissions. Or because the author has been there, but not to others.

In Ethan Mollick’s book “Co-Intelligence: Living and Working with AI,” the author writes that the greatest benefits from using AI in these broad use cases are for those who have the least knowledge of the topic. An example was a student business idea project, where, after starting to use AI, they had much more original ideas than students who didn’t use it. However, comparing this among professionals—AI’s creativity lags behind what experts in the field do. But field professionals who use AI assistants for certain tasks are already pulling ahead of those who don’t. So—use it, but wisely!

A couple of Artificial Intelligence tools worth considering for travel planning:

• Gemini, ChatGPT—for planning days and eateries. You can ask for data in CSV format and upload it to Google My Maps so it is immediately available as a route.
• Gondola—for redeeming loyalty points for hotels.