Three 20-year-old hikers on Mount Shasta in northern California found themselves in distress after a series of poor choices. They ran short of food, over-estimated their stamina, and ignored the late-day warning to turn back, reaching the summit at 7:30 p.m. and beginning a night descent. Their planning relied heavily on Google’s Gemini chatbot, and by morning they were lost on a difficult section, one with a bruised knee.
A separate Mount Shasta visitor set up camp beside a “No Camping” sign, insisting AI advice gave her permission to stay. Park officials, identified as Meyers, also reported a man at a trailhead holding an AI-generated map that misplaced major landmarks on the opposite side of a nearby peak. Both cases illustrate how algorithmic recommendations can conflict with on-ground regulations.
In British Columbia, a helicopter rescued a 16-year-old who had slipped onto a ledge on the Widowmaker Arete of Crown Mountain. The teenager disclosed that he had used Anthropic’s Claude chatbot to outline his itinerary, though he also consulted Google Maps and the AllTrails hiking app. The rescue drew media attention across Canada and the United States.
Whether AI will raise the overall number of backcountry incidents remains uncertain. Meyers’ office, which manages a large wilderness region in northern California, typically conducts about twelve rescues annually, and the current year appears to align with that average. In contrast, Wyoming’s search-and-rescue calls have fallen, with 312 reports so far in 2026 compared with 466 in 2025 and 377 in 2024. No clear causal link to AI has been established.
Veterans of rescue operations note that chatbots are the newest in a line of digital tools that have occasionally misdirected outdoor enthusiasts. About ten years ago, hikers turned to Instagram and other social platforms for inspiration, sometimes with fatal outcomes. Earlier still, reliance on Google Maps or constant smartphone connectivity produced its own set of navigation errors. Jim Moss, a Colorado recreation attorney and volunteer rescuer, says media pushes have simply increased outdoor participation, leaving the impact of robotic advice unclear.
Researchers highlight two specific dangers of over-reliance on conversational AI. Clemson University professor Chris Flathmann describes a phenomenon he calls “cognitive surrender,” where users stop questioning the system and accept its output uncritically. Chatbots can also “hallucinate,” presenting fabricated or outdated information from sources such as Reddit threads or old YouTube videos as authoritative guidance. Flathmann advises users to verify any route or safety data against official public-land agency maps, trail association sites, guidebooks, or direct consultation with park rangers or experienced guides.