Back to Articles
‘Don’t make my mistake’: Aussie traveller’s border nightmare after relying on ChatGPT

Nine.com.au

ENRICHED

Details

Date Published
4 Aug 2026
Priority Score
1
Australian
Yes
Created
4 Aug 2026, 04:02 am

Authors (1)

Description

Max and his family were left scrambling at an African border.

Summary

The article details a case of AI hallucinations and misinformation where an Australian traveller received incorrect visa advice from ChatGPT, leading to border entry complications in Africa. It highlights the inherent risks of relying on Large Language Models for high-stakes administrative or legal information, such as immigration requirements. While not addressing existential risks, it underscores the safety and reliability concerns of frontier AI models in real-world applications and mentions how AI can facilitate more sophisticated social engineering and scams. The content serves as a practical warning for Australian citizens regarding the current limitations of AI accuracy in navigating global governance systems.

Body

sharesShare articleWith artificial intelligence apps fast becoming more common, it can be tempting to use them to get travel advice.However, many holidaymakers are finding out the hard way they’re not always accurate.Max Phelps relied on AI to sort out visas and got the wrong information. Supplied/Max PhelpsAdvertisementWhen Max Phelps, 58, from Sydney was planning a once-in-a-lifetime trip to Africa, he logged onto popular AI program, ChatGPT.But the advice it gave left the money coach and mortgage broker, who was holidaying with girlfriend, Julie and grandson, Matthew, scrambling.“My girlfriend and I knew that we wanted to see Victoria Falls and turned to AI for the practicalities, like the best places to see it from, where to stay, and where to fly into,” he told nine.com.au.With Artificial Intelligence apps fast becoming more common it can be tempting to use them to get travel advice. iStockRead moreIs there any snow on Aussie ski fields this season? We find out“AI recommended seeing Victoria Falls from both the Zimbabwean and Zambian sides, then a safari in Botswana’s Chobe National Park. “So we built our itinerary around that.”ChatGPT told him to get the KAZA Univisa, which it claimed covered travel across all three countries with unlimited re-entry for 30 days. AdvertisementHe made the mistake of not cross-checking this information.Victoria Falls at the border between Zambia and Zimbabwe iStockAs Aussies, they didn’t need a visa for Botswana or Zambia, only Zimbabwe. And the visa they got wasn’t correct.“The problem with the KAZA Univisa is that while it allows unlimited travel between Zimbabwe and Zambia, it becomes invalid if you leave those two countries for one night or more,” he said.AdvertisementAdvertisement“It wasn’t until we tried to re-enter Zimbabwe after the safari that border officials told us our visa was no longer valid.”Luckily, they were able to buy a visa at the border. But Max said it took time, caused stress and cost $43 each in visa fees, on top of the $72 each they’d paid.Max Phelps relied on AI to sort out visas and got the wrong information. Supplied/Max PhelpsAdvertisementRead more‘Behind a hill was the most incredible hotel room I’ve ever stayed in’He warned other travellers not to make the same mistake.“If you’re using AI to plan a trip, don’t just take its first answer as the complete picture, especially for anything involving visas, borders, or legal requirements,” he said.Research from Send My Bag found nearly four in 10 Australians have used AI to research or plan a trip, with adoption highest among Millennials and Gen Z.AdvertisementAdvertisementAnd while over half trust AI to find the best prices and create itineraries, only 35 per cent trust it with visa or immigration requirements.Founder Adam Ewart admitted it can “take the pain” out of research. However, it makes mistakes, he warned.“It can only provide answers based on the information it has,” he told nine.com.au.“The best way to use AI for travel is to treat it as a starting point, not a replacement for official travel advice.”AdvertisementGetting the right visa is one of the basics of travel. AAPSometimes AI even makes up places entirely.Last year, dozens of tourists flocked to a Tasmanian town in search of picturesque hot springs they’d seen online, only to learn that they’d been duped by an AI-generated image.Experts at RMIT University in Melbourne said you need to put your detective’s cap on.AdvertisementAdvertisementRead moreStrict new Bali visa rules crack down on influencers and content creators“Spotting a fake isn’t about finding one giveaway – it’s about building a picture,”  Dr TJ Thomson from the School of Media and Communication said.He told holidaymakers to “look closely for inconsistencies or glitches” and use multiple sources of evidence.“If you’re not sure, do a reverse image search on Google,” he said.AdvertisementInternet security company Mcafee has also warned just how much information you are giving away if you post a holiday photo publicly.Posting photos from holidays publicly can make it easier for scammers. Getty Images/iStockphotoSome AI models have a more than 90 per cent accuracy rate at detecting the location a photo was taken based on the visuals in the photo alone, it said.And that makes it easy for scammers.AdvertisementAdvertisement“AI removes the guesswork, allowing attackers to build highly specific, contextual scams at scale,” it warned.That could include sending believable context around an attack, such as naming the city or country where the victim had been, fooling them into thinking a message or email is legitimate.The in-flight habit that Aussies hate more than rowdy childrenAussies are ditching Thailand for this unlikely Asian destination‘Don’t make my mistake’: Aussie traveller’s border nightmare after relying on ChatGPTVirgin drops $429 return Bali flights, plus this month’s best deals