From wrong hours to dangerous hikes: the hidden risks of letting AI plan your holiday

Your next holiday itinerary might have been written by an algorithm that's never actually left the house.

While more than one in ten Australians are now using AI to dream up their getaways, experts warn that these "digital travel agents" come with a catch.

From hallucinated opening hours to potentially life-threatening hiking advice, the line between a seamless trip and a travel disaster has never been thinner.

A recent Compare the Market survey found more than one in 10 Australians have used AI to look for destination recommendations.

People are also using AI to find deals, recreation activities and accommodation, to create itineraries, find flights and transportation, and understand currency conversion, the study found.

With more people using AI to research travel, Australian Travel Industry Association (ATIA) CEO, Dean Long, says it's important to understand what these tools can and can't reliably deliver.

"AI can be a smart starting point for inspiration but travellers need to fact-check critical details and use trusted sources when making booking decisions," Long tells Explore.

"But, it also leaves you at risk of fraud and the very real chance of disappointment of missing out on those key attractions you'd always wanted to see because details like how far ahead to book are missing or wrong."

Long says AI is already the new "front door" for dreaming about a holiday.

"But let's be clear: AI isn't a travel professional; it's a search engine with a better vocabulary," he says.

"It's changing the 'research' phase but it doesn't change the 'reality' phase. The problem is that AI is always looking in the rear-vision mirror; it can only tell you what has already happened, been reviewed, or tagged a thousand times before. It's a reflection of the past, not a map of the now or the next destination waiting to be discovered."

Southern Cross University adjunct professor in tourism, Anne Hardy, says AI "has already changed the way we travel".

"AI has had a profound influence on how we travel, where we travel to and how we plan our travel," she says.

Examples include suggesting itineraries, showing prices and transport options, as well as for navigation and support while travelling such as language translation, and to assess risk including crisis or weather. It is also being used to monitor queues, Hardy says.

But she warns users to be wary as "it is estimated that 90 per cent of AI itinerary suggestions have at least one inaccuracy in them."

"This may be an inaccurate opening hour, or it could be far more serious such as suggesting tourists walk through dangerous suburbs, or hike in remote areas when they are under prepared. So as consumers, we need to treat AI as an idea generator, but not the stand alone expert," Hardy says.

Her primary concern with AI is when travellers use it when making travel decisions that involve risk and the outdoors.

"When you ask AI to suggest hikes in Tasmania, there are several hikes that it describes as being easy or moderate when in fact they are not. It can also underestimate what safety equipment people should take with them, as well as the likely weather conditions," she says.

"We have had record numbers of people die or need emergency assistance over the past 12 months in Tasmania. Research is urgently needed to understand the role that AI has made in their decisions that have ultimately led to their need for emergency help or, worse, their deaths in our natural areas."

"Perhaps in the future, instead of having a human tour guide, we might use an AI bot as a tour guide. When we check into a hotel, instead of calling the receptionist for information and advice, we might be able to use an AI bot on the hotel room TV," he says.

Griffith University - Griffith Institute for Tourism professor of practice, Daniel Gschwind, says there is a "potential flip-side to all the AI advancements".

"Travel at its core is about discovery, curiosity, the unfamiliar. I ask myself, will AI take away the wonder, the surprise, the delight of unexpected discoveries? Is that the price we pay for wall-to-wall information and hyper-planning?" he says.

"AI will know all your past preferences and previous choices and guide you accordingly. Will our world still hold exciting secrets to discover?"