A DontBeThatTourist Story
When AI Sends You to a Place That Doesn’t Exist (Literally)
AI travel planning is sending tourists hiking toward destinations that don’t exist. Here’s what happened, why it keeps happening, and how to not be that tourist.
Don’t let a chatbot become your Sherpa. AI predicts what should come next — it doesn’t verify facts the way you think it does. Same confident tone, whether it’s true or totally made up.
DontBeThatTourist Lesson #37: Verify Before You Fly
The “Sacred Canyon of Humantay” That Never Was
Pharmacists double-check prescriptions. Travelers? We should double-check directions. Especially when those directions come from artificial intelligence that can’t always tell the difference between a hidden gem and a hallucination.
So here’s another “What Could Possibly Go Wrong?” story: tourists hiking toward destinations that don’t exist.
Two tourists in Peru proudly showed a local trekking guide their AI-generated itinerary. They were headed to the “Sacred Canyon of Humantay,” complete with poetic descriptions, spiritual vibes, and the nearly $160 they reportedly paid to get there.
Small problem. There is no Sacred Canyon of Humantay. AI had mashed together two real names — Humantay Lake and the Sacred Valley — into one glorious hallucination.
And while this story is funny in hindsight, it’s also dangerous. Peru’s Andes climb above 4,000 meters, where weather changes fast and phone signals vanish.
Related reading: They Said Flip-Flops Were Fine… They Were Wrong — my Bali waterfall hike disaster. Read it here
The Pattern
When AI Hallucinates, You’re the Punchline
More people asking robots for directions, more robots inventing the destination.
Sources: Greetwell & Deloitte, Sep 2026
AI travel planning is incredible at generating ideas, travel inspo, and captions that sound way better than “me standing next to a rock.” But here’s the catch: AI doesn’t verify facts the way you think it does — it predicts what should come next. It writes confidently… even when it’s wrong.
That’s how travelers have been:
- Given marathon itineraries through Italy that required Olympic-level teleportation.
- Told by ChatGPT that the last ropeway down Japan’s Mount Misen left at 5:30 p.m. It didn’t. A couple reached the top just in time to watch it close.
Case Files
Real-Life AI Oopsies That Deserve Their Own Reality Show
Because misery loves company, here are a few of my favorite “you can’t make this up” AI travel fails. Fair warning: these are widely reported stories, not ones I witnessed firsthand — which, in a post about AI making things up, feels like the responsible thing to say out loud. 😅
Case 01
The Visa That Vanished
Spanish creator Mery Caldass asked ChatGPT whether she needed a visa for Puerto Rico. It confidently said no — and never mentioned the ESTA authorization she actually needed. She missed her flight.
(The chatbot wasn’t exactly wrong about the visa. It just skipped the part that gets you on the plane.)
Case 02
The Phantom Cable Car
An elderly Malaysian couple drove over 300 kilometers from Kuala Lumpur to the town of Kuak Hulu after watching a viral TikTok promoting a scenic cable car ride called “Kuak SkyRide.”
They checked into their hotel and asked the staff where to find it.
Nobody knew what they were talking about.
Because Kuak SkyRide didn’t exist.
The polished travel video was AI-generated. The scenery was fake. The happy tourists were fake. Even the “reporter” presenting the segment was AI-generated. The woman reportedly said, “Why would anyone want to lie? There was even a reporter” — and wanted to sue the presenter, until the hotel employee explained the presenter was AI-generated too.
The disappointment? Extremely real.
Case 03
When the Official-Looking Website Is the Problem
Here’s the one that should make every careful traveler a little nervous: a tourism site promoting Fukuoka, Japan published AI-generated articles about attractions that didn’t exist — including the “Koga sashimi set,” which it named as a popular local dish.
There is no Koga sashimi set. I checked. Now I just want sashimi.
Let that sink in. Not a chatbot. Not a TikTok. A tourism website backed by official municipal endorsements — the exact kind of place you’d go to verify the other stories on this list. It apologized and pulled the articles six days after launch. Six days.
And those are just the ones that made the news.
The Fine Print
Why This Keeps Happening (And It’s Not One “Bad” AI)
Here’s the important part most people miss: this isn’t a “bad app” problem. It’s a “how the thing is built” problem. These systems are designed to be helpful, imaginative, and confident. (Notice what’s missing from that list? “Accurate.” Obviously.)
The tricky part: the made-up stuff doesn’t sound unsure. It arrives in the exact same confident tone as the true stuff — same adjectives, same certainty. That’s the trap.
That’s why:
- General chatbots invent “hidden gems” (translation: they made it up, but make it ✨mystical✨)
- Itinerary generators stitch together half-real routes like a toddler with safety scissors and a glue stick
- AI travel reels fabricate entire attractions — cable cars, viewpoints, vibes
- Even official-looking tourism sites have published places that never existed. A Fukuoka promo site apologized and pulled its AI-written articles six days after launch. Six days.
Real talk: the more creative the output, the higher the hallucination risk. Which is exactly backwards from how most of us use these tools — we ask for the most creative stuff (“hidden gems! secret spots!”) and get the least reliable answers in return.
💊 Travel Rx: How to Use AI Without Getting Used by It
AI can be an excellent starting point. Just don’t let it become your only source of truth. Think of it like taking travel advice from that one friend who says, “I swear I read this somewhere.”
The good news? Most travelers already use it the smart way: 43% use AI travel planning for research and inspiration, and 34% for building itineraries. (Adobe/eMarketer) Brainstorming buddy? Yes. Booking agent? Absolutely not.
Here’s your AI → IRL Double-Check Dose:
- Map it twice. Two apps. Satellite view. If one shows a trail and the other shows cows, trust the cows.
- Find the source. Official sites for hours, ropeways, ferries, and permits. Not “travelblog.fun.”
- Reality-check elevation. A “leisurely 4,000-meter walk” is not a stroll. That’s a climb.
- Google the local name. AI mixes words like a tourist at a tapas bar. Locals will tell you if your canyon exists.
- Sort reviews by date. If the last update is from 2019, something changed — usually the schedule.
- Have a bailout plan. Offline maps, battery backup, and a turn-around time even if the view looks Instagrammable.
✈️ Traveling soon? I’m putting the finishing touches on a free Layover Guide — the carry-on version of traveling smarter, not harder. Join the email list and you’ll be first to get it when it drops.
Red Flags Your Itinerary Is Fiction
- Adjectives: 10. Coordinates: 0.
- “Secret spot locals don’t want you to find.” (Translation: it doesn’t exist.)
- “Perfect sunset timing!” (Translation: you’re sleeping on that mountain.)
- Place name sounds like a remix: “Sacred Valley of Humantay Rainbow Canyon Falls.” No. Absolutely not.
AI Can’t Replace Experience
I use AI travel planning tools all the time — they help me brainstorm captions, summarize travel news, and research destinations. But when it comes to actual trip planning, I rely on verified sources, my network, and the behind-the-scenes data only travel agents see: supplier alerts, closures, regional safety updates, and yes — whether that “secret lagoon” is real or just well-written fiction.
So when in doubt: Ask a human who’s been there. We’ll make sure your canyon exists — and that you make it back down before sunset.
The Mic Drop
The Part Where I Drop the Mic (Gently — It’s Expensive)
Here’s the thing nobody selling you an AI travel planning tool wants to say out loud:
Your gut already knows this, and the data agrees: only 5% of travelers book exactly what AI suggested. The other 95%? They go check — 59% head straight to Google. But Google can’t tell you whether the canyon exists either. That’s literally my job. As one travel study put it: “AI starts the trip. Trust closes it.” (Mower, Sep 2026)
A chatbot has never stood at 4,000 meters wondering where the canyon went. I’ve stood in enough “wait, this definitely isn’t what the website showed” moments to know the difference between a plan and a plan.
I don’t just hand you an itinerary I hand you the version where someone already asked the annoying questions.
AI is a fantastic brainstorming buddy. But when it’s time to actually book the flights, cross the border, and climb the mountain? You want a travel advisor in your corner. Preferably one who’s already made the mistakes so you don’t have to.
(Hi. It’s me. I’ve made so many.)
So use the robots for AI trip planning ideas. Then come talk to a travel advisor before you go.
Curious what it’s like having an actual travel advisor plan your trip — one who double-checks the canyon exists?
Travel smarter. Laugh harder. Explore deeper. Live kinder. 🩷
— Chelsea
Sources
- Peru “Sacred Canyon of Humantay” & Mount Misen ropeway: BBC Travel, “The perils of letting AI plan your next trip” (Sep 2025)
- Mery Caldass / ESTA story: BuzzFeed (Aug 2025)
- Malaysia “Kuak SkyRide” cable car: PetaPixel (Jul 2025)
- Fukuoka AI-generated tourism articles: The Mainichi
- 1-in-6 AI hallucinations + Deloitte growth stat: Greetwell/Stacker (Sep 2026)
- How travelers use AI (43% research/inspiration, 34% itineraries): eMarketer, Adobe data (Sep 2026)
- “AI starts the trip. Trust closes it.” (5% book what AI suggested): Mower (Sep 2026)