Road less travelled: Escape algorithm, travel differently with AI
Paris, Venice, Santorini, Barcelona – the usual suspects. AI algorithms tend to nudge you toward the same places because they’re trained on what’s most visible online. The illusion of personalised advice makes people less likely to question it — and that’s how AI risks intensifying overtourism
Ask an AI service like ChatGPT or Google Gemini to suggest a destination for your next summer holiday, and it will happily produce a list of attractive spots. But most will sound familiar.
Paris, Venice, Santorini, Barcelona – the usual suspects. AI algorithms tend to nudge you toward the same places because they’re trained on what’s most visible online. The illusion of personalised advice makes people less likely to question it — and that’s how AI risks intensifying overtourism.
The use of AI for travel planning is booming. A recent survey found usage has doubled in a year, with younger travellers leading the way. Nearly one in five Britons aged 25–34 now rely on AI tools to plan their trips.
In my own research, I analysed ChatGPT’s travel suggestions and found it gravitates to the world’s most-visited destinations by default. Lesser-known or more sustainable locations only appear when travellers explicitly request them.
This could easily worsen overtourism, which already strains residents in popular areas. In Mallorca, locals are demanding limits on flights and holiday rentals. Venice has even introduced a day-tripper fee to manage visitor numbers.
AI will add to that pressure if millions of people follow the same algorithmic trail. These systems are trained on data that amplifies visibility — reviews, blogs, hashtags — and inevitably reinforce what’s already popular. Accepting the defaults means more of the same, and more stress for places already under siege.
Yet travellers aren’t powerless. With a bit more intent, AI can still uncover fresh and fascinating destinations.
The key is to ask better questions. Generic prompts like “best beaches in Europe” or “beautiful cities” produce identical results. Try instead: “Which towns are reachable by train but overlooked in most guides?” or “Where can I go in July that’s not a major tourist hotspot?” Push the system, follow up, and scroll past the first few results — that’s often where the surprises lie.
Timing matters too. AI tends to focus on peak season because that’s when most reviews and travel content appear. Asking about off-peak months — say, the Italian lakes in October or the Greek islands in May — helps sidestep this bias.
You can also prompt AI to dig deeper into its sources. Because it relies heavily on English-language content, it favours international hot spots. But it can also surface local blogs and tourism cooperatives. Try “Spanish-language blogs about Asturias” or “community-run agritourism in Slovenia” to discover something genuinely off the beaten path.
AI can even help compare costs, timings and carbon footprints — if you’re willing to look past the surface. After all, these systems are built to serve what’s popular, not what’s sustainable. (Though the same technology could easily be tuned to prioritise rail over air, or highlight locally run businesses first.)
AI’s convenience is seductive but also predictable. If your holiday plans could be copy-pasted from Instagram, any real sense of adventure is lost.
Use AI as a starting point, not the final word. Guidebooks, local media and conversations with residents restore the unpredictability that makes travel memorable. By asking sharper questions, shifting timing, and seeking local voices, travellers can turn AI from a tool of congestion into one of discovery.
The next time you ask ChatGPT where to go, make it work harder — or settle for the same crowded itinerary as everyone else.
The Conversation
Road less travelled: Escape algorithm, travel differently with AI
