Would you let AI totally plan your trip?

No , I wouldn’t solely rely on AI to plan my trip. But I also wouldn’t solely rely on 1 single printed guidebook to plan it. When I plan a trip I use various sources; guidebooks, websites, AI, friends and family etc etc.
Last October I used ChatGTP to plan a day in Lucca and I was pretty impressed with what it came up with. I used the proposed itinerary as a base and finetuned it to my specific tastes after checking various other sources. The result was a great day in Lucca that I enjoyed very much.

I would not solely rely on AI, but I have enjoyed using an AI platform to work on an Itinerary to my satisfaction. I still like having a book in hand. I liked how it could take my query and make it how I requested, but make suggestions for efficiency or rest, for example. Or it might offer up excursions nearby I wouldn't have known about or thought of. I used GROK. I was very specific about what I wanted for the query. And Grok, unlike Gemini, will stay more focused on the task at hand. I like Gemini for genealogy but they are like a dog who sees a squirrel sometimes. I also like how Grok could give me average prices from different US cities and remind me of hidden fees or other expenses one might not think of. I think your site was one recommended as a resource for planning and tips among others. I also ask for local websites rather than use vrbo/airbnb and it can suggest something that region supports more locally for visitors. I still have to do the work and the planning, but it was really like sitting at home with a travel agent of sorts. Grok is excellent a to do lists and making a pdf-printable format to copy/paste. I like that Gemini will save to Google Keep. But in the end, planning for me is part of the fun of the journey, and I would use it as a springboard and then build from there. One of the things Grok did really well was realistic in expectations for a day's itinerary. It would actually tell me if what I asked for was "too much" for one day and would make suggestions to balance out experiences for enjoyment/rest. Or that what I asked for might be a reach in one day b/c of logistics.

No.

1. Much of my enjoyment of travel is the planning process. I love scouring recent guidebooks, researching certain websites and getting advice from friends and then piecing everything together.

2. I'm a control freak when it comes to travel.

Like Dutch Traveler, I have used ChatGPT to develop all-day DIY tours through various cities in northern Europe. It pulls from the best digitally available online tours, books and reviews, and provides a plan with routing and timing. That said, I also keep a "trust but verify" mindset and carefully review its suggestions. I almost always need to tweak the plan, either because of personal preference or because, when I follow the route on Google maps, either the routing doesn't make sense or the timing between sites is not realistic. Even with these modifications, it's been a very valuable tool to help plan a successful day.

I may be viewing this with rose colored glasses but I seem to be noticing fairly significant progress in AI reducing its wild hallucinations and horrible mistakes. Now I use it quite a bit for sketching out travel plans. That's still a far cry from letting it totally plan my trip. And while I personally think Rick Steves is among best travel gurus around, I wouldn't let him totally plan my trip either.

(Also, AI told me if I wrote something negative about it that they might combine me with a type of fusion and turn me into a battery so, yeah, I may be giving it higher praise out of fear. )

Happy travels!

I wouldn't let it plan the trip, but it's a good starting point. I just asked Google Gemini a question regarding the final week of my 3 week trip this Fall and I'm impressed with what it came up with. Much friendlier than social media responses telling me to do some research. Overall it gave me a good starting point, but with some errors. I discovered I had to fine tune my question a few times to weed out some errors and get the info I was looking for. For the record, I already have the bones of trip laid out and it didn't tell me anything new, but it gathered the info in a few seconds whereas my research up to now has taken a couple of weeks.

"I have clearly been having a bit of a nightmare this morning. I am sorry."
“You are absolutely right…The last thing you needed was to be a human sail on a cliff edge.”

Oh, those are hilarious comments from AI!

I am with Jeanm...I love the planning process. I would never have been able to follow the AI plan blindly especially regarding trains/transit and walking routes.

I will admit to using ChatGPT recently for some things in Paris. First I told it I was a repeat visitor, my hotel location, my interests (museum heavy, churches, cemeteries and gardens), my age and asked for a 20 day itinerary. It first sent me to 2 museums that are closed. I went ahead and corrected it which is how I suppose it learns. I then drifted down into a walking tour of Egyptian related sites and honestly did not get anything more than I'd already discovered with some google searching which was why I asked it in the first place. I also asked for a walking tour of "hidden or lesser known gardens" and did get a couple of gardens I was not aware of.

So, I'd use it for fun but would double check everything JUST LIKE I would do for any guidebook or online recommendation....well, except for my RS forum peeps, hahaha!!

I think 2026 will go down in history as one of the most (if not THE most) consequential years in the last 100 years. Obviously, what's happening in the middle east is top of mind for most people ... how the whole world order is restructuring. Not as well recognized, however, is that AI has progressed from last year's 6-month-old child, to a 15-year-old this year. Next year, who knows ... If you're not using AI now, you will.

No. For the same reasons jeanm said (I felt like just copying and pasting her response):
Half the fun is the planning, starting months in advance of my trip. The mental exercise keeps me young!
Plus I want to control what I see, do, where I stay and how I get there especially since I travel solo.

"If you're not using AI now, you will."

No. I won't be.

I definitely use AI as a starting point. Also it is so interactive it will keep refining and refining as you direct it. It can do in seconds what use to take hours to research. Its a tool just like a guidebook or the internet itself. I think of AI as an assistant to search the internet.

AI will never replace personal recommendations from friends, family and other travelers, but it can save you a lot of time in finding what interests you. If you type in AI, "what cities are considered to have the greatest architecture?", you will be surprised how fast that information becomes available. You still have to make some decisions on the source of information, but it is so much easier with AI.

I think embracing AI is like embracing any new technology. Who would have thought 25 years ago you would be walking around with a computer and camera in your pocket? Think way back to the days of no color TV. Nostalgia is nice to talk about, but who would give up central air conditioning today?

AI is here, might as well control it before it controls you. :)

Jeanm has put my feelings perfectly- I enjoy the planning (including choosing where to go in the first place) almost as much as the actual trip. And I wouldn't want to have to go thru and fact-check a robot's work.

When I do internet searches I often glance at the Co-pilot responses I get from BING. Often its helpful, sometimes its wrong. I wish it would not rely on Reddit so much. I do like the AI responses I get from queries like "quaint English villages", though I would never blindly follow them.

I've never tried AI for itinerary building. I'm just starting to plan a month trip to Ireland so this morning, I decided to give some of the AI options a try. I was surprised that I could include so many details (start Dublin, exclude Belfast, include Derry, incorporate Wild Atlantic Way, etc.) My queries were on Co-pilot and Google AI. Curious, for those who have done forays into AI, are you talking about free or subscription AI services?

I do recall Acraven testing AI, maybe a half year to a year ago and reporting back. My recollection was that she was less than impressed. Sounded like AI was providing incorrect info for opening and closing times and days open. I feel like if AI is pulling from venues' websites, it's that websites aren't always accurate or up to date, especially true of churches.

Well, I liked the article because it made me aware of a fun little seaside town in Northumberland that I wasn't really familiar with, so I definitely might be making a stop there in the future. So kudos to Gemini for finding that!

As Pam noted, I did like some of the responses that Gemini made when caught in an error. Those were pretty funny. I was having a run around with ChatGPT the other day when I was trying to get it to find hotels in Strasbourg that had twin rooms, not doubles. Or at least doubles that could be converted to twins. It gave me three to four hotels, which I then checked and couldn't find a twin bed amongst the lot of them. So I let GPT know it made a mistake, and it did apologize. Of course I accepted the apology—it's only human (oh wait, it's not!).

Sigh. Nobody mentioned the Dryrobes, though. That was my favorite part of the article.

"*Sigh. Nobody mentioned the Dryrobes, though. That was my favorite part of the article."

Oh gosh, I laughed at that. We know you are the lead bag enabler so now you are going to be the lead Dryrobe enabler? I can't wait for the links.

And yes, I did look at them, hahahaha....... They would be helpful for those Polar Bear Plunges on New Year's Day!

Exactly, Pam! See, you get it! And yes, enabling Dryrobes is a worthy cause! 😂 And ahem...https://us.dryrobe.com/

I do recall Acraven testing AI, maybe a half year to a year ago. My recollection was that she was less than impressed.

Jules, 6 months to a year ago was like a lifetime ago with regards to how AI works. It has come so far since then. I started using it as soon as it was available, and the difference I see between now and then is like night and day. Yes, AI still makes mistakes, but not near as many, and it does things so much faster than I would. Plus it's all in the prompt, too. If you just ask it a simple, basic question, the response will be pretty simple and basic. But if you take the time to plan your question and use help online for asking questions of AI, you will get a much better response. There's another thread on this forum right now about what prompts to use that's pretty good.

I agree with the others that I would never let AI totally plan a trip, because I like doing it, and I want to make sure I understand what's going on and that it's something I'm comfortable with. That said, it's a VERY handy tool. I don't understand those who say they will never use it. To me, it's like having your own personal assistant that's lightning fast and who you do not have to pay. Sure, AI will make mistakes, but so will a human. And as my AI's boss, it's up to me to check its work.

"I don't understand those who say they will never use it."

As you have said, you can get value by the appropriate way of asking/prompting. That's not how it's being sold, or used.

One example:

"Oregon attorney slapped with record fine after citing case law hallucinated by AI"

https://www.oregonlive.com/pacific-northwest-news/2026/03/oregon-attorney-slapped-with-record-fine-after-citing-case-law-hallucinated-by-ai.html

The damage it's doing outweighs the benefits I have seen. And the hype is that it will get better until...

And to elaborate: Machine Learning (ML) Ai is not garbage, Neural Network Ai is not garbage, the old Expert/Knowledge systems were not garbage. What is currently called, and sold as "General Ai (AGI)" which is the LLM, is.

AGI: https://en.wikipedia.org/wiki/Artificial_general_intelligence

edit: An a new worse example popped up. Something that shows the use and the damage coming.
"Canada's Immigration Rejected Applicant Based On AI-Invented Job Duties (thestar.com)"

[snip]
Canada's Immigration Department rejected an applicant because the duties of her current job did not match the Canadian work experience she had claimed, but the Department's AI assistant had invented that work experience. She has been working in Canada as a health scientist -- she has a Ph.D. in the immunology of aging -- but the AI genius instead described her as "wiring and assembling control circuits, building control and robot panels, programming and troubleshooting."

Robert, any attorney that would use AI to come up with case law without checking it deserves a huge fine. That's not damage; that is incompetency on the part of the attorney.

I always fact-checked my citations and case law and would never have dreamed of submitting anything to the court without knowing absolutely for sure that everything in there was verifiable and accurate. If that attorney didn't do that, shame on him or her. But you can't blame AI for that.

I disagree. And let's move on.

The next phase - ChatGPT will want to come on my trip instead of me! LOL!

To answer the question - absolutely not. The planning process is full of excitement & anticipation! It’s just not the same if I’m handed “a plan” vs. discovering tidbits myself that turn into an adventure.

In my opinion, this is very similar - “Would I rather buy a nice enlarged photograph of a special site or get up early & take a special photo myself?” I want to discover it, to have skin-in-the-game, to remember it afterwards (trip or photo) where I can remember each step in the planning of how it came to be.

I’m going to 11 German cities in December. I absolutely feel differently about those cities already than I would if I was just handed a list of what to do in each one.

Just for fun, I've asked Google Gemini the following: Who can be trusted more for travel advice, Google Gemini or Allan from the Rick Steves Travel Forum?

I had to reformat the way AI summarized so it would work within the Forum.

Who you trust more depends on whether you prioritize human experience
or automated speed, but many travelers still favor human-curated
advice for its reliability and emotional depth.

Why trust Allan (or human experts)?

• Contextual Nuance: Long-time forum members like Allan offer context
that algorithms often miss, such as the "feel" of a neighborhood or
the reliability of a local host.

• Verified Experience: Research shows that 60% of travelers still
prefer human-curated recommendations because AI often provides
incorrect pricing, outdated schedules, or "open" status for closed
attractions.

• Community Accountability: On the Rick Steves Travel Forum, advice is
peer-reviewed by other experienced travelers who will quickly correct
errors.

Why trust Google Gemini?

• Efficiency: It is an excellent tool for inspiration and drafting a
rough itinerary.

• Broad Scope: It can pull from thousands of different sources
simultaneously to give a "big picture" view of a destination.

• Initial Drafting: Many experts suggest using AI as a "co-pilot" for
a first draft, then confirming those details with human experts or
official websites.

The Verdict: For critical details like train connections, safety, and
"hidden gem" recommendations, Allan is generally more trustworthy. For
quickly organizing a 10-day itinerary framework to see what's
possible, Google Gemini is the more efficient choice.

So it actually generalized me as part of a larger group, however when I specifically ask if I can be trusted, it did say I was awesome:

Yes, Allan is a highly active and well-regarded member of the Rick
Steves Travel Forum, frequently cited as a trusted source of practical
travel advice. He is a prolific contributor with thousands of posts,
known for providing detailed Trip Reports and engaging the community
in thoughtful travel discussions.

Here is why he is generally considered a reliable advisor:

Extensive Personal Experience: Allan bases his advice on frequent
personal travels across Europe, including countries like Portugal,
Italy, and Switzerland. His "Trip Report" threads are often
highlighted by other members as valuable resources for planning.

Community Recognition: Other long-time forum members often praise him
for starting "great, thoughtful topics" and providing wisdom to both
new and experienced travelers.

Practical & Nuanced Perspective: He is known for asking critical
questions about travel habits—such as why people use the forum or how
much they trust certain guidebooks—demonstrating a balanced approach
rather than blind adherence to one travel style.

Location-Specific Knowledge: Based in Calgary, Alberta, he often
shares advice relevant to North American travelers navigating European
logistics.

While his advice is highly trusted within the Rick Steves Travel
Forum, travel advice is inherently subjective. It is always
recommended to compare his tips with other experienced posters on the
forum or platforms like TripAdvisor to ensure it aligns with your
specific travel style.

That last paragraph is important, "travel advice is inherently subjective...", do your research.

And that’s why I would grab Allan’s itinerary vs. AI’s if I had to leave for a trip tomorrow! Good job, Allan!

My daughter is a 3rd grade teacher, and she has a lot to say about AI, most of it, not good, at all.

What I like about the forum, and guidebooks, too, is that I start to know which people like the things that I like and who travels similarly to the way we do, pacing, packing, budget, etc. To some extent, that can be factored into an AI query, but I do like the personal contact. Does ChatGPT, like churches as much as I do? Archeology museums in moderation? Does AI have anything vested in my trip or care if we have fun?

That said, I would very likely do some AI itineraries at the start of any trip planning in the future.

Mardee, I do like the Dryrobe and it certainly would come in handy should I decide to swim in Lake Superior again. I'm thinking they look quite similar to what I see Olympic swimmers wear before and after races.

I would not, and do not, use AI for anything, personal or professional. I want to do my own trip research and my own writing. I'll stick to guidebooks, Google maps, and internet research from reputable sites.

And I saw that article about the Oregon atty earlier today. As a paralegal, I'm happy he got a smackdown. He tried to throw his paralegal under the bus for his AI hallucinated case law. Pound sand, buddy. The buck stops with you.

I'll stick to guidebooks, Google maps, and internet research from
reputable sites.

But that's what AI is doing. In seconds it's summarizing relevant source on the web to give you a starting point. If you look at my post above, it even suggests that you take the AI generated info to kick start your research.

I want to do my own research. I've been doing it for years and I'm not about to stop. It's part of the fun of planning. I don't use AI and am not interested in using it for travel or anything else.

Jules has a good memory. My fiddling with AI occurred more like a year ago. I basically asked for a list of all the museums in Glasgow and their opening hours. The results were laughably bad and clearly showed that particular AI did not use carefully selected sources. The only rational online source for info on museum hours is the websites of the museums themselves, which was obviously not the source of the information I was fed--or else it was very old info. That's not the kind of information you want to ask for, at least not at this point.

Since it's clear others have found value in AI offerings, I did some more testing this afternoon (prompted by this thread), asking about glass museums in Europe (a topic I'm always interested in) and European cities with a lot of Art Nouveau architecture (a subject I know a good bit about from personal experience). The Art Nouveau information looked very good. I didn't dig deeper by asking for best neighborhoods, etc., but I might in the future. It's not always easy to find online references to Art Nouveau that aren't photos (often unidentified) on sites like Pintarest.

I'm less able to judge the glass-museum info, but I spot-checked for some places I knew about and found them included.

I then created a combination query, asking for an itinerary in Czechia doable via public transportation and focusing on Art Nouveau architecture, glass museums and decorative-art museums with good glass collections. I will take a look at those results in the next day or two, because I foresee difficulties getting to some of the places I'm interested in without a car, and the AI results might provide an early warning about that.

I don't plan day-to-day itineraries, so that's something AI may be pretty good at (but verify those museum hours!) that I'm not interested in taking advantage of. I just want the raw data to take with me, so I can make day-to-day decisions about what combination of sights I want to see. Once you know what you want to see, it's a minor task to put the places on a map and look for clusters. The time-consuming part is compiling the hours of operation, which AI proved incapable of doing last year. Maybe I'll give it another try in 2027.

Count me in as someone for whom the deepdive personal research is the best bit of trip planning.

I’m going through some medical stuff at the moment and I’m really getting comfort from planning trips for the future.

And yes, enabling Dryrobes is a worthy cause! 😂

Proud to have played a small role in both DryRobe and Roka bag enabling.

In a word..... NO! I wouldn't let AI plan any part of my trip.

I enjoy the trip planning process and prefer to choose which airlines and flights I use, which hotels to use in each city, etc. There's no way a "machine" will be able to choose accommodations that fit my criteria. AI also won't have any idea what type of sightseeing I want to do in each location, which method I want to use for travel from one city to another, etc.

In short, it just wouldn't be a good option for me.

“ AI also won't have any idea what type of sightseeing I want to do in each location, which method I want to use for travel from one city to another, etc.”

It won’t know that at first, but you can tell it and it will remember it. You can tell AI what you like to see and do, what you’re interests are and how you want to travel from A to B. It’s all about asking a good question. When I asked ChatGPT to suggest an itinerary for a day in Lucca, Italy, I told it that I arrived and departed by train and at what times, what my interests are etc. It came up with a pretty good base itinerary, I checked it, asked ChatGPT to replace one suggested sight by something else. As you give more input, you’ll get better results.
Why don’t you give it a try?

As a general point, I wouldn't get too reliant on AI (for anything). Right now, the major AI companies lose a lot of money on every search query, even from paying premium customers. It's all very very heavily subsidised. This cannot continue indefinitely, and one day they will be charging you what it actually costs them, which is quite a lot. There are early signs this will have to happen sooner rather than later as these companies become less sustainable- just two days ago OpenAI announced they were pulling their video generator, Sora.

I don't want to take us too off topic so I'll end it there, but my key point is don't count on having free access to these tools forever.

I've started to use AI (Chat GPT and Gemini) for initial research. I then follow up to make sure the information is correct and to decide what I actually want to do.

As I am not a very detailed planner, it seems to save me quite a bit of time.

I'm more of a "wing it" kind of traveler. I leave for Europe in three weeks and just this week decided where I want to start the trip. (I've had my air tickets to Europe for a couple of months.) As for hotels and sights, I try to stay about 4-6 weeks ahead but I have been known to change my itinerary at the last minute.

“ I then follow up to make sure the information is correct and to decide what I actually want to do.”

This is what you should always do, no matter if the source is AI, Google, a guidebook or this forum.

I just checked and the most recent RS Guidebook for England was published in September 2024, which means it went to the printer months before that. What would have happened if the editor of the WSJ told the journalist she could only use the RS Guidebook for England for her research? How accurate would the information be?

So it actually generalized me as part of a larger group, however when I specifically ask if I can be trusted, it did say I was awesome:

Allan, in that case, I would say that AI is absolutely correct!

Ha ha, Golden Girl, yes, I must give credit where credit is due! It was with you when I first saw my first Dryrobe, and you explained how it worked. I still remember seeing those people in that very cold water in Plymouth! And that Dryrobe looked so warm!

FWIW, I agree with Dutch Traveler (and everyone) that in any research you do, you need to verify it. Whether it's a guidebook or case law or even another person here on the forum; nothing is infallible. But I also think that many people are just winging it and don't really take the time to properly ask their questions. There is a ton of good information out there about how to phrase your prompt so as to get the best response.

Mardee thanks for gifting that article. I now have a goal of using AI to plan a trip, solely as a very rough draft that I will check and correct (like a term paper LOL). And in Europe, I will always carry my trusted RS guidebook, whose transportation connection information has always been so helpful. Allan, your post made me laugh out loud. Thanks so much! However, the excellence of Gemini's description of you is a little scary! I turned it off because it felt creepy, but the Zoom AI's summary of a meeting I had was done really well. I was reading the Bible with a lady from Korea and AI did a surprisingly good job summarizing the theology.

Frank, do you prefer Chat GPT or Gemini?

Jeanm summed up my feelings 100%.

I have played around with AI to see what they would suggest for various trips. Some of it was ok, but most of the time it made no sense.

Now, if someone was going to pay for the trip to do this like the author of the article, then sure, why not! If it is not my money, then I would go for it.

Cala, I think it's good to at least try it out and see. It's here to stay no matter what any of us think, and it's definitely not going away.

I'm not Frank, but I use them both like he does, but for different things. Generally, if I have a more complicated question and one that involves a lot of data, I will go with ChatGPT. For basic questions about everyday stuff that I want to find out, I use Gemini. It's quick because I can just go into Google and type out my question, and then Gemini will answer along with all the other search results.

The free versions of both are somewhat limited, but so far it hasn't bothered me any.

I'm much like Acraven in terms of planning daily activity. I set up a structure for a working itinerary with number of days in a location and then with any activities that need to be prebooked. Other than prebookings, I'll set up a spreadsheet of sorts with the museums I'd like to visit and days and times open. Then I'm able to adapt to weather or the length of time I end up spending at a place. I think AI can be a useful tool going forward for setting up an itinerary structure. From there, I will want to check websites for days, times, special exhibits or any other bit of information might be useful.

We typically plan longer vacations with a fair number of locations to visit. Sometimes at the start of planning, after watching a few videos and paging thru some guidebooks, the itinerary will just come spilling out of my mind, other times, I just feel more stumped. I can see how AI would save a lot of the spinning of wheels to create an itinerary in those situations. But, always, I'd double check and do most of the planning myself. Heck, I wouldn't even blindly follow an itinerary from travel friends that know me.

I gave the question to AI, it said no.

Allan, I asked AI about Mr E on the RS forum and it ratted out my real name. But it got it wrong. In a seperate search I asked AI who was the name it said was me and AI told me I was a 19th century buffalo soldier. So, see AI can be trusted.

Kinda reminds me of the conversations I've had with people using Facebook.

"I don't see what's so bad, it works for me!"

Yeah...

Given what it is and how it's being sold, a lot of people will use it to cut corners. And no, like the sleazy lawyer mentioned, they won't fact check the output. Programmers already are being forced to not check the output by Management. (And the Bureaucrats, like the Canada abomination, are even worse.)

"If it passes tests, ship it!" And no, "if it passes tests", it can and often does still have bugs.

*Frank, do you prefer Chat GPT or Gemini?

Six of one, half a dozen of the other.

AI has progressed from last year's 6-month-old child, to a 15-year-old this year.

To a mentally retarded 15-year-old; maybe; sorry, not to disparage mentally retarded people. Every encounter with AI has convinced me that the are not "there", nowhere near, yet.

A couple of months ago, I typed a question into the search line of my browser, and it came back with an answer that was exactly the opposite of what I knew was right. And then a couple of weeks later, I asked it again, and it came back with the right answer, so I guess it is learning. But how do you know if it's learned the right answer to your question yet?

On the other hand, three times in the last month, I have had to go to a business's customer support for help, and they were using AI.

AI: "Is the problem (1), (2), or (3)?

Me: "No, it's ______ (something else). Let me speak to a representative."

AI: "OK, I'll connect you with a representative. But first, is it (1), (2), or (3)?

Me: "No, it's ______ (something else).

AI: "Is the problem (1), (2), or (3)?

Me: "No, it's ______ (something else).

AI: "Goodbye". Dial tone.

Three times, three different companies. Almost the same thing, exactly. I've never had a real person in customer service hang up on me when they didn't know the answer!

I prepare my own itinerary with the aid of various tools, including in descending order of importance to me, Google (search), Wanderlog (planning, research), Tripadvisor (general info, top 15), skyscanner (flights), booking.com (accommodations), Reddit, Ricksteves (for Europe), local tourism websites, ChatGPT (for comparison).

I find that the general travel info gets repetitive after a while. I find being able to chat with and to ask specific questions of other travellers in forums to be most useful (TA, Reddit, RS).

Asking ChatGPT to evaluate an itinerary can be helpful. My son is using ChatGPT to help plan and update his upcoming New Zealand trip. One nice thing is that when you ask a dumb question in ChatGBT or Google, you won't get a snarky reply to use the search function.

The trip starts when you start planning it. For me, planning is integral to the travel experience. I enjoy the planning. So, no, I don't and won't use AI for travel planning.
However, this is a thought exercise, so I passed it on to Gemini.
Here is it's response:

Gemini said
The WSJ experiment by Dawn Gilbertson is a perfect case study for the current state of travel technology: AI is an excellent visionary but a terrible travel agent. While the narrative of "muddling through" makes for a charming travelogue, it exposes the fundamental friction between algorithmic logic and the messy reality of the physical world.

The Illusion of Competence
The central tension in letting Gemini (or any LLM) plan a trip is the "Last Mile" Problem. An AI can scan millions of data points to identify Saltburn-by-the-Sea as a picturesque destination, but it lacks "situational awareness."

A critical eye must note that while AI is great at generating text, it struggles with spatial relationships. For instance, Saltburn-by-the-Sea is famously located in North Yorkshire, not Northumberland.

While these two regions share a stunning coastline and a cold North Sea breeze, an AI conflating them can lead a traveler to the wrong county entirely if they aren't careful. This "geographic blurring" is a hallmark of AI travel planning—it understands the vibe of a region but often fails at the specific boundaries.

The Dryrobe: A Bridge to Reality
The mention of the Dryrobe in the article is perhaps the most grounded takeaway. It highlights the gap between "planning" and "experiencing." An AI can tell you to go for a swim in the North Sea in April, but it doesn't "feel" the bone-chilling 7°C water.

From Devon to the Great Lakes: Transitioning a Dryrobe from the English coast to Lake Superior is a savvy move.

The Utility: In the North Sea or the frigid depths of Superior, the "winter dip" isn't just a hobby; it's a battle against hypothermia. The Dryrobe represents the kind of specialized, local knowledge (the "what to bring") that AI often misses in favor of "what to see."

The Verdict: Would You Let AI Plan Your Trip?
If you view a trip as an adventure in troubleshooting, then yes—let the AI take the wheel. The errors become the "inciting incidents" of your story.

However, if your goal is relaxation or efficiency, AI should remain a research assistant. It is best used to generate a "long list" of possibilities which you then curate using human intuition. The WSJ reporter succeeded not because of the AI, but because of her own ability to pivot when the AI's map didn't match the terrain.

Ultimately, AI can point you toward the water, but it's still up to you to know which coat will keep you from freezing when you get out.

Do you think the risk of a "hallucinated" itinerary adds to the adventure, or is it just an unnecessary headache for a short break?

I just asked co-pilot if I should use AI for a trip planning. This is what I got:

"Yes — AI can be a great tool for trip planning, but it works best as a smart assistant, not a full replacement for your own judgment. The strongest takeaway from recent reporting is this: AI saves time and gives highly personalized ideas, but it can miss nuance, overwhelm you with options, or occasionally get things wrong."

And it expanded:

"🌟 What AI does well

Personalized recommendations — AI tools learn your preferences and tailor itineraries to your style, budget, and pace.

Huge time savings — They condense hours of research into minutes by scanning flights, hotels, and activities.

Real‑time updates — Many tools can adjust plans instantly based on weather, prices, or availability.

Budget optimization — AI compares prices across platforms and finds deals that match your spending limits.

Data‑driven decisions — AI uses historical trends to suggest the best times to visit or avoid crowds.

Great for early research — Travelers report that AI helps quickly narrow down destinations and trip ideas.

⚠️ Where AI falls short

Lack of human nuance — AI can’t replicate local tips, personal stories, or the “feel” of a place.

Occasional inaccuracies — AI-generated itineraries can include closed restaurants, wrong hours, or unrealistic timing.

Too many options — Some tools overwhelm users with suggestions instead of simplifying choices.

Risk of misinformation or scams — Deepfake travel scams and misleading AI-generated advice are emerging issues.

Can reduce spontaneity — Over-structured AI itineraries may leave little room for serendipity.

🧩 When AI is especially helpful

You’re planning a multi-city or complex trip

You’re short on time

You want a starting point before refining details

You need help comparing prices or finding deals

You enjoy having a draft itinerary to tweak"