Tourists Trusted AI Route Advice, Got Stranded in the Tatras, TOPR Rescued Them

Two Lithuanian tourists got stranded in Poland's Tatra Mountains after trusting a route suggested by artificial intelligence. Hoping to shortcut their way to the Valley of Five Polish Lakes (Dolina Pięciu Stawów Polskich), they followed a chatbot's directions across Świnicka Ławka, a passage that leads into difficult climbing terrain they couldn't retreat from. A helicopter from the Tatra Volunteer Search and Rescue service (TOPR) had to be sent out.

The pair of tourists had set out to reach one of the most popular destinations in the Polish Tatras, the Valley of Five Polish Lakes, using suggestions generated by artificial intelligence to shorten their hike. Instead of a safe tourist trail, the system suggested crossing Świnicka Ławka, a stretch of terrain that requires climbing experience, protective equipment, and knowledge of current conditions, none of which any AI tool could assess in real time.

A day of back-to-back rescues

The AI-related incident wasn't the only rescue operation that day. TOPR logged a real string of accidents on Saturday, July 4, responding to calls 25 times. Rescuers also dealt with two climbers trapped on Mięguszowiecki Szczyt, a scientist with a broken leg who had to be extracted from the Lodowa Cave in Ciemniak, and a hiker who got stuck among fallen trees and had to survive the night using a survival kit.

The surge of calls at a single moment made logistics difficult for rescuers, especially since the helicopter couldn't be deployed to every incident at once. TOPR described the day as a textbook case of accidents clustering together, with several independent incidents overlapping and forcing rescuers to prioritize which operation to run first.

Once again we are calling for people to choose their mountain destinations sensibly and not overestimate their own abilities - TOPR statement

AI with no knowledge of the terrain

The Tatra case illustrates a specific limitation of AI tools when used to plan mountain routes. A chatbot can suggest the shortest path based on maps and descriptions, but it doesn't assess current weather conditions, trail conditions, the presence of snow or ice, or the preparedness of the person asking. Unlike a mountain guide or an experienced hiker, the system won't warn that the suggested route requires a rope, crampons, or prior climbing training.

Mountain rescuers have long stressed the need to plan routes according to one's own skills, but the growing popularity of chatbots as a first source of information before heading into the mountains adds a new element of risk. A route can look credible in a text description and still lead somewhere impossible to return from without specialized equipment.

What it means for tourists

The case concerns the Polish mountains directly, but the problem has a broader dimension. As more people ask AI assistants for trip plans, restaurant picks, or sports routes, the number of situations where a system's lack of context and up-to-date local knowledge translates into a real threat to health and life keeps growing. TOPR and other rescue services will likely need to account for this factor more often in their incident statistics.

For travelers, there's one practical takeaway: AI suggestions for mountain routes are best treated as a starting point for further checking, not a ready-made plan, and should always be cross-checked against official maps, current TOPR advisories, and one's own assessment of the terrain's difficulty.

Sources: Two tourists got stranded in the Tatras. Their guide was... artificial intelligence (rmf24.pl), Artificial intelligence led tourists into a trap. TOPR intervened (forsal.pl)