
AI (Artificial Intelligence) is here and in a big way. And there is no doubt that it can be extremely valuable. But it has significant limitations, especially when trying to understand and then plan an expedition, safari, or luxury travel experience. In fact, over just the past few weeks, there have been some great examples of AI’s limitations and the confusion and frustration it causes for both consumers and travel businesses. But first, let’s take a step back.
How Good is the Data AI is Using?
In the June 2026 issue of Consumer Reports, the magazine rated cruise lines, and, to be honest, its findings were baffling. It stated, “We rated 19 cruise lines based on a CR Member survey of 18,850 people who took a cruise in the last 36 months,” and the result was that National Geographic – Lindblad was by far the best cruise line (9 points or over 10% higher) over all others. While it is a solid product, according to Consumer Reports, it is tops almost every category while being more expensive than Silversea, Regent Seven Seas, Seabourn, and 35% more expensive than Viking, with Silversea and Viking offering expeditions as well as cruises.
It shows National Geographic-Lindblad rated 5 out of 5 for “Food Quality and Variety”, higher than Silversea, Celebrity, Oceania, and more. No disrespect, but one of the biggest complaints my clients have about National Geographic-Lindblad – besides it being extraordinarily expensive for what you receive – is the food (which now you must place your dinner order at breakfast!). The same goes for its staterooms (5 of 5): they are small and pretty barren, and most are on old ships. Another thing is that Consumer Reports left National Geographic-Lindblad’s excellent youth programming out of ratings, probably because of the lack of data, not that it doesn’t exist. Confused yet about whether you could or should rely on Consumer Reports for your cruise decision?
As an experiment, I asked both ChatGPT and Claude the following question: “How does National Geographic-Lindblad rate compared to other cruise lines?” and both referenced Consumer Reports right at the top (and, honestly, how many people actually read the entirety of what follows?). I do note that both ChatGPT and Claude pushed me in two different directions, with one emphasizing that NG-L isn’t top-tier for suite quality or “dining polish,” and the other trying to have me keep it to expedition lines. (I also have to note the heavy reliance on TripAdvisor and Cruise Critic, both of which heavily rely upon untrained guests who probably have had one experience, could be cheerleaders or upset passengers, and you have no idea if they have any expertise or are within your demographic (Applebee’s or Fine Dining, for example).
Now, put yourself in the position of trying to figure out what the heck National Geographic-Lindblad is and how it really is positioned for what really matters: Your Travel…if you even know or can even decipher what it is that actually matters!
AI has a Built-In Prejudice to Agree and Will Make Things Up if it has to.
And, of course, there is the other problem: Both ChatGPT and Claude are programmed to agree with you so that you will engage with them more often. A recent study showed that Claude agrees with you at least 49% of the time. So, if you write, “I heard Tanzania is the best place for seeing the Big 5,” chances are it will guide you down that road, as opposed to having a discussion about whether ticking them off is your goal or having a truly immersive African experience. And, of course, if there are other places that are equally as good or better. (There are!)
Further, if you know me, you know that the last thing I want is an AI agent that doesn’t tell me what I am thinking is wrong or that I have missed something. I want to know these things so that the best travel experience is had, not to stroke my ego.
Worse, AI agents make things up, called an “AI Hallucination”. I have had it make up hotels (including fake reviews) and logistics that are impossible. In the legal community, courts are sanctioning lawyers for citing AI hallucinated law cases that don’t exist, and interpretations of cases that are actually contrary to what an actual case stated.

Real World Examples and Why AI Isn’t as Helpful as You May Think!
In my world, I recently had a client come to me with an itinerary of sorts for a safari in Tanzania, suggested by their friend. Why Tanzania? I think you know the answer: AI suggested it. And while it can be a great experience, I suggested Botswana would be better. However, the “expert” AI mode was already in the client’s head. Making matters worse, this itinerary did not account for the logistics of getting from A to B, let alone whether there was too much or too little time at each safari lodge. Then there was the clear focus on luxury lodges rather than the safari experience, so the lodges, while amazing, were not where they really wanted to be. Oh, and then there was the budget. Yeah, that blew up the whole thing. So much wasted time for me and so much frustration for the clients.
I also recently had another client who wanted to go to the Arctic. Well, it was pretty much the same thing, except this time it was more about challenging my decisions. When I was a bit emphatic that I was the expert and that AI pulls from who knows where, she said, “Well, don’t be so hard on my Chat Buddy. He recommended you!”
I am also aware of technology companies looking to enter the travel market as a “one-stop shop” from air to transfers to tours to hotel to cruise; even giving some options for the cruise or hotel. Sounds pretty interesting, huh? Well, in the premium and luxury market, it is not as easy as finding a good fare, but I have to consider a particular airline, time of day, layovers, class of service, seats, etc. , not just day and cost. Oh, then there is how long it takes to be in City A, and what kind of transfer (some folks want a black car, others need a van for luggage, others are happy with a taxi or Uber). Hotels? The top agents stay in various hotels so they can guide each client to the best proven hotel options, not a computer screen. And, of course, Goldring Travel has added amenities at hundreds of hotels, and some clients insist on getting those Marriott points!
Of course, I haven’t even ventured into what my clients’ interests are. I mean, a tour including churches is great for some and not so much for others. (The phrase, “ABC – Another Bloody Church,” was not made up by me.) And what about an amazing culinary experience, whether it be a food tour, a cooking class, or dining (street food or Michelin-starred)?

How to Effectively Use AI
AI is a wonderful tool. But that is all it is: A tool.
I use AI every day, be it for spell and grammar checking (boy, does it drive me crazy trying to change the meaning of sentences to what it “thinks” I want to say), having my Making Waves articles able to be listened to in my voice (and, soon, podcasts), and as a research tool for simple things as to what cruise lines will be sailing in Norway in November (which it gets close enough, but not 100% accurate) to more complex travel questions. But AI is nothing more than an add-on or starting point.
However, when it comes to travel, what I hear the most of is, “I’ve been doing all this research and, honestly, it has gotten me so confused.” The good thing is that it’s when people call me! I have been to Antarctica nine times and, without question, I know a lot more about Antarctica than whatever article(s) it is pulling from. And because I understand my clients, I can quickly sort through MY database to provide the information, with detailed explanations and photographs, that is most useful to my clients. And I can get very specific.
The same holds true if you want to go on a cruise or spend time in Barcelona, Istanbul, Greece, or wherever.
So, please feel free to use AI as a tool, but no more than that. Yes, it will get better over time. How much better? Over how long a time period? Those are big questions. And, as I have discussed, don’t you have enough travel questions you want answered?