Field Cases and the Phenomenon of Cognitive Surrender
According to a WIRED article, three 20-year-olds became stranded on a rock face on Mount Shasta in northern California after asking for help: they had brought insufficient food, overestimated their physical fitness, reached the summit at 7:30 p.m., and begun the descent in the dark. As they later admitted, their poor planning was partly due to relying on Google’s Gemini chatbot. Nick Meyers, head of the Mount Shasta Wilderness Climbing and Avalanche Center program at the U.S. Forest Service, commented that these hikers had at least tried to inform themselves, but in his view “they switched off their brains.”
The phenomenon is not isolated: according to Meyers, another visitor pitched her tent near a prohibition sign, claiming that AI had told her she could do so, while a man wandered along a trail with an AI-generated map that placed important features on the wrong side of the mountain. In Canada, a helicopter crew rescued a 16-year-old from a ledge on the Widowmaker Arete on Crown Mountain: the boy had used Anthropic’s Claude chatbot to plan the trip, though he had also consulted Google Maps and the AllTrails app.
Chris Flathmann, a professor of human-centered computing at Clemson University, explains the mechanism: chatbots can “hallucinate,” that is, present false information with confidence, and their use can lead to so-called “cognitive surrender,” in which the user stops reasoning critically. If AI pretends to be an expert on a subject the user does not know, the user lacks the tools to challenge it, and returning to critical thinking in a dangerous situation is difficult. Moreover, chatbots tend to say what the user wants to hear, in an authoritative tone, even when the sources are Reddit threads, dated YouTube videos, or invented content.
Numbers and Caveats: What Is Confirmed and What Is Missing
Chip Jenkins, superintendent of Grand Teton National Park in Wyoming, reports a sharp increase over the past year in the percentage of visitors who say they use AI for travel information: from zero to 26 percent, a curve that in his view will not flatten out. However, it has not been demonstrated that growing AI use leads to more accidents and rescues: Meyers’s office, which covers a vast wilderness area in northern California, records an average of a dozen interventions a year, and 2026 appears to be average. In Wyoming, rescue calls have decreased compared with 2025: 312 so far in 2026, versus 466 in 2025 and 377 in 2024.
Jim Moss, a Colorado lawyer specializing in recreational activities and a volunteer in rescue operations, observes that chatbots are only the latest tool offering new ways to get into trouble: ten years ago social media pushed people to discover destinations, and before that Google Maps and smartphone connectivity created problems. What remains to be understood, he adds, is how a robot’s questionable advice will affect all these people. Flathmann highlights a difference: TikTok and Instagram inspire, whereas AI provides instructions, right or wrong, and encouragement; it does not seek to deceive, but it is the natural evolution of a tool of this kind.
The Experts’ Recommendation
The advice from rescuers and experts is the same regardless of the tool used: verify information against authoritative and up-to-date sources, such as maps and websites of the agencies that manage public lands, websites of trail associations or guides, and even better, speak with a guide, a forest officer, or a ranger. Meyers’s conclusion: anyone going into places that can kill must take planning seriously.
The picture that emerges from the article is therefore twofold: on the one hand, documented cases of hikers who attributed wrong decisions to AI and a reported sharp increase in use in a national park; on the other, the current absence of data linking increased chatbot use to an increase in accidents and rescue operations. The distinction between operators’ perceptions, the protagonists’ accounts, and the available statistics remains central to assessing the phenomenon.
