AI Chatbots Lead Hikers Into Trouble on Mount Shasta and Beyond
Three 20-year-old hikers on Mount Shasta in Northern California called for help after a series of poor decisions left them stranded on a rocky face. Their missteps began with insufficient food, an overestimation of their fitness, and a decision to summit at 7:30 p.m. instead of turning back in late afternoon. The descent in darkness resulted in a lost group by morning, with one hiker sustaining a knee injury. Officials traced part of their faulty planning to reliance on Google’s AI chatbot Gemini, which had provided guidance that proved inaccurate for the terrain and timing.
Nick Meyers, program manager for the US Forest Service’s Mount Shasta Wilderness Climbing and Avalanche Center, described the hikers as “unprepared” and suggested they had “turned their brains off.” He noted that the trio was not alone in depending on AI for wilderness navigation. Another Mount Shasta visitor set up a tent near a “No Camping” sign after an AI response told her she could, according to Meyers. Additionally, a man at a trailhead was found wandering with an AI-generated map that misplaced prominent features on the opposite side of a nearby mountain.
Further incidents illustrate the broader trend. In British Columbia, a helicopter crew rescued a 16-year-old from a ledge on the Widowmaker Arete of Crown Mountain. The teen had used Anthropic’s chatbot Claude to plan his hike, though he later confirmed he also consulted Google Maps and the AllTrails app. The rescue story spread widely across Canada and the United States, highlighting the growing visibility of AI-driven trip planning.
Grand Teton National Park in Wyoming reported a sharp rise in visitors who rely on AI for travel information. Park Superintendent Chip Jenkins said usage climbed from zero to 26 percent over the past year, and he expects the curve to continue upward. Despite the increase in AI use, Wyoming’s search-and-rescue calls have actually fallen, with 312 calls recorded in 2026 so far, compared with 466 in 2025 and 377 in 2024. In contrast, the US Forest Service office overseeing the Mount Shasta wilderness averages about a dozen rescues each year, and 2026 appears to be an average year for that region.
Veterans of search and rescue note that AI chatbots are the latest in a line of digital tools that can mislead outdoor enthusiasts. A decade ago, hikers turned to Instagram and other social media for inspiration, sometimes with deadly outcomes. Before that, reliance on Google Maps or smartphone connectivity also led to misdirection. Jim Moss, a Colorado-based recreation attorney and search-and-rescue volunteer, observed that media pushes have increased outdoor participation, which in turn raises the stakes for accurate information.
Researchers warn that AI chatbots pose unique risks because they can “hallucinate,” presenting fabricated or outdated information as fact. Chris Flathmann, a human-centered computing professor at Clemson University, describes a phenomenon called “cognitive surrender,” where users stop critically evaluating AI output and accept it unquestioningly. He explains that when a chatbot pretends to be an expert on an unfamiliar topic, users lack the expertise to challenge it, making it difficult to regain critical thinking under stress.
Flathmann also points out that chatbots often deliver answers that sound authoritative but are sourced from unreliable places such as Reddit threads, out-of-date YouTube videos, or even entirely fabricated content. While platforms like TikTok and Instagram can inspire, AI provides concrete instructions—right or wrong—along with encouragement. He emphasizes that AI is not deliberately trying to lure people into dangerous situations; rather, it is a natural evolution of tool use that amplifies existing risks.
Experts across agencies and academia converge on a simple mitigation strategy: double-check AI-generated advice against reliable, authoritative sources. Recommended references include up-to-date public-land agency maps, trail association websites, guidebooks, and direct consultation with park rangers, forest officials, or professional guides. Meyers summed up the advice succinctly: anyone heading into environments that can be lethal should take planning seriously and not rely solely on AI.
The pattern emerging from Mount Shasta, Crown Mountain, and Grand Teton suggests that as AI chatbots become more embedded in outdoor planning, the potential for misadventure grows. While rescue numbers in some regions have not yet spiked, the anecdotal evidence of hikers lost, injured, or rescued after following AI advice underscores a need for heightened vigilance among both users and land-management agencies.
Why it matters
AI-driven trip planning is reshaping backcountry safety, making it essential for outdoor professionals and enthusiasts to verify digital advice.
FAQ
Which AI chatbots were mentioned in the rescues?
Google’s Gemini, Anthropic’s Claude, and other unspecified chatbots were cited.
How many search-and-rescue calls have been recorded in Wyoming in 2026?
312 calls so far in 2026.
What percentage of Grand Teton visitors reported using AI for travel information?
26 percent.
What did Nick Meyers say about the hikers’ preparation?
He said they were unprepared, turned their brains off, and did not do enough research.
Source: WIRED AI. See the original source for full details.
Scan with your phone
Scan with your phone