The Pulse
AI gets people into trouble in the backcountry
Hikers have used chatbots while planning trips that ended in wilderness rescues, but the reported cases do not establish that AI use is increasing accidents. Experts warn that confident-sounding chatbot advice can weaken critical thinking a
AI.info Team · Written with AI · fact-checked ·
At Grand Teton National Park, the share of visitors who said they used AI for travel information rose from zero to 26 percent over the past year, superintendent Chip Jenkins told High Country News. The figure anchors a November 1 WIRED story about hikers turning to chatbots for backcountry plans—and the risks when those answers sound more dependable than they are. The cases include people who needed rescue, but experts say the reports do not establish that AI use is driving a rise in accidents.
Three hikers summited Mount Shasta at 7:30 p.m.
In the most detailed case, three 20-year-old hikers on California’s Mount Shasta relied partly on Google’s Gemini while planning their climb. They brought too little food and overestimated their fitness. Although they had not reached the summit by late afternoon, they continued and summited at 7:30 p.m., then began descending in darkness. By morning, they were lost on rough terrain and one hiker had hurt his knee.
Nick Meyers, U.S. Forest Service program manager of the Mount Shasta Wilderness Climbing and Avalanche Center, said the group had made an effort to research the trip, but added: “I think they kind of turned their brains off.” The account describes a mix of poor preparation and decisions on the mountain; it does not attribute every mistake to the chatbot.
Other incidents complicate the chatbot story
Meyers also described a visitor who pitched a tent beside a “No Camping” sign and argued that AI had told her camping was permitted. His team encountered another person at a trailhead with an AI-generated map that placed prominent features on the wrong side of a nearby mountain. The article does not identify which chatbot produced that map.
In British Columbia, a helicopter crew rescued a 16-year-old from a ledge on Crown Mountain after he used Anthropic’s Claude to help plan his trip. The hiker later said he had also consulted Google Maps and AllTrails. That detail matters: the account does not show that a chatbot alone sent him onto the ledge.
Rescue figures offer no simple verdict
The rise in reported AI use at Grand Teton does not, by itself, show that more people are getting into trouble. Meyers’ office averages about a dozen rescues a year, and he said 2026 was shaping up to be an average year. Wyoming recorded 312 search-and-rescue calls so far in 2026, compared with 466 in 2025 and 377 in 2024, according to the figures cited by High Country News.
Search-and-rescue volunteers have seen other technologies create hazards, too, from social-media posts that draw people to difficult places to mapping apps and lost phone connectivity. The available figures do not yet connect increased chatbot use with an increase in backcountry accidents or rescues. The cases instead point to a specific concern: hikers may treat polished answers as reliable route and safety instructions, even when they cannot judge the advice themselves.
Fluent directions can outlast a bad answer
Chris Flathmann, a human-centered computing professor at Clemson University who studies AI and search and rescue, described the risk as “cognitive surrender.” “When humans work with AI systems, they can get so deep into it they give up thinking critically for themselves,” he said. A chatbot may present information from unreliable sources—or fabricated details—in the same assured tone as sound guidance, leaving an inexperienced user with little basis to spot the difference.
Experts quoted by High Country News recommend checking plans against current public-land agency maps and websites, trail-association resources or guidebooks, and asking a ranger or local guide when possible. Nick Meyers, U.S. Forest Service program manager of the Mount Shasta Wilderness Climbing and Avalanche Center, put the stakes plainly:
“If you’re going to go into places that will kill you,” Meyers said, “you should take your planning seriously.”
The unresolved question is whether the rise in chatbot use will show up in rescue data; for now, officials say the numbers do not establish that trend.