In partnership with
It’s New Year's Eve!
Hello, Curse and Coffee friends,
Today, we explore ChatGPT’s travel blunders.
Hit reply and let us know what you think (we read all of your kind words).
Coffee at the ready…
The Big Sip
The take: AI can generate a travel itinerary in seconds. It just can't guarantee any of it exists.
What happened: CNN Travel journalists tested ChatGPT's planning capabilities across five major cities — Atlanta, Hong Kong, New York, London, and Bangkok — and the results included closed venues, fabricated locations, and routes that ignored basic geography.
Why it matters: In Hong Kong, ChatGPT suggested visiting the airport as a tourist destination. In New York, it scheduled the 9/11 Memorial at 8:30 pm (an hour and a half after it closes). A separate study found ChatGPT recommended "Antico Caffe Ponit" in Rome — a café that doesn't exist. The tool hallucinates venues with the same confidence it recommends real ones.
What to watch: OpenAI recently opened ChatGPT to third-party travel integrations with Booking.com and Expedia. Early testing has encountered technical issues and reliability concerns. The tool is being monetised before the errors are fixed.
ChatGPT itself admits it should serve as "a starting point" rather than replacing "real research." At least someone's reading the fine print.
Sponsor Break
Before we slurp into today’s brew…
Here are some wordies from today’s sponsor.
Winning “Brewery of the Year” Was Just Step One
Coveting the crown’s one thing. Turning it into an empire’s another. So Westbound & Down didn’t blink after winning Brewery of the Year at the 2025 Great American Beer Festival. They began their next phase. Already Colorado’s most-awarded brewery, distribution’s grown 2,800% since 2019, including a Whole Foods retail partnership. And after this latest title, they’ll quadruple distribution by 2028. Become an early-stage investor today.
This is a paid advertisement for Westbound & Down’s Regulation CF Offering. Please read the offering circular at https://invest.westboundanddown.com/
Here’s Your Brew
The problem is the confidence.
Ask for a London pub, and it invents The Crown & Anchor. Ask for a museum, and it sends you to one that closed two years ago. Ask for transport to an island, and it ignores the ocean.
Each recommendation arrives formatted identically, with the same authoritative tone, whether the information is accurate, outdated, or entirely fictional.
This is the hallucination problem (and travel planning is where it hits hardest). ChatGPT doesn't verify. It predicts.
It generates the next most likely word based on training data, which means a closed restaurant and an open one look the same in the output.
The CNN test found some wins.
One New York tester discovered the Stavros Niarchos Foundation Library — a genuinely helpful, uncrowded alternative. A Bangkok tester found lesser-known neighbourhoods in Thonburi after multiple rounds of refinement.
But both required extensive prompting and verification.
The 90% error rate from the SEO Travel study tells the real story. Nine out of ten itineraries contain mistakes.
The tool creates a lot of homework.
Two Sides, One Mug
→ Pro-AI: ChatGPT can rapidly synthesize information and generate starting points faster than manual research. With detailed prompts and verification, it surfaces options you might not find otherwise.
→ Anti-AI: A tool that's wrong 90% of the time is a liability. Travelers without local knowledge can't spot the errors until they're standing outside a closed museum.
→ Our read: AI travel planning works best for people who already know enough to catch the mistakes. A beta test with the public as QA.
Receipt of the Day
Source: SEO Travel, "90% of AI Travel Itineraries Are Inaccurate," September 2024
"Chat-GPT suggested a New York itinerary that included dinner at one permanently closed restaurant one night and another one the next!"
Why it matters: The SEO Travel study tested 100 two-day itineraries across ten cities. Beyond the 90% error rate, it found ChatGPT recommended places that don't exist — like "Antico Caffe Ponit" in Rome. The tool hallucinates venues with the same confidence it recommends real ones.
Spit Take
"90% error rate. 100% confidence. That's the product."
Your Coffee Break Links (and water cooler chatter)
📎 The full CNN investigation: Five cities, five journalists, dozens of errors. The London and Hong Kong tests are ruthless. Worth reading in full. (CNN Travel)
📎 The 90% error rate, explained: SEO Travel tested 100 itineraries across ten cities. They found closed museums, invented cafés, and two permanently closed restaurants in a single New York dinner plan. (SEO Travel)
📎 The visa disaster: In March 2025, author Mark Pollard trusted ChatGPT's claim that Australians didn't need a visa for Chile. They have since 2019. His conference was cancelled. His Instagram post got 15 million views. (CamJon Travel)
Join your team of caffeinated skeptics.
Opinionated world news that respects your time.
One bold take, the best counter, and the receipt(s) that prove it (all in sixish minutes).
Mugshot Poll 📊
Would you trust ChatGPT to plan a trip?
You can read all our back issue newsletters for free here.
For the love of coffee, see you on Friday!
We’ll be travelling tomorrow (so no newsletter). Have a fantastic New Year's Eve/Day!
I appreciate your support in 2025.
See you all in 2026!
Enjoy your celebrations…
And keep them caffeinated.
How did we do?
Thanks for reading!
Are you subscribing?
Be sure to get your daily curse and coffee fix by hitting the button below.