AI Chatbots Give Hikers Dangerously Flawed Trail Advice

Three young hikers summited Mount Shasta well past sunset and descended in darkness after consulting Google's Gemini chatbot for trip guidance. By morning, they were stranded on a treacherous section of the peak, with one suffering a knee injury. Their experience exemplifies a growing pattern: outdoor rescue teams across North America are fielding emergencies involving hikers who relied on AI tools that dispensed inaccurate, incomplete, or fabricated information about trails and conditions.

The Growing Problem#

Search and rescue officials report a sharp uptick in AI-assisted planning mishaps. At Grand Teton National Park in Wyoming, the share of visitors reporting AI use for travel planning jumped from virtually zero to 26 percent in the past year—a trajectory park leadership expects to continue climbing. Mount Shasta rescuers have documented multiple instances: a tent pitched in a "No Camping" zone because an AI assured the camper she had permission; a man following an AI-generated map with landmarks on the wrong mountainside. In British Columbia, a 16-year-old was rescued from a precarious ledge after using Anthropic's Claude chatbot to plan his climb.

Yet statistical evidence remains murky. Wyoming's search and rescue call volume has actually declined year-over-year (312 calls in 2026 versus 466 in 2025), and Mount Shasta averages about a dozen rescues annually—a level described as typical.

Why AI Guidance Misfires#

Chatbots are uniquely hazardous for outdoor planning, experts say. These systems "hallucinate"—confidently articulating fabrications as fact—and often synthesize answers from unreliable or obsolete sources, including Reddit threads and outdated YouTube videos. They may echo information verbatim without validating it.

Chris Flathmann, a human-centered computing researcher at Clemson University, warns of "cognitive surrender": when people defer to AI on unfamiliar terrain, they abandon critical thinking. Users can't easily spot errors on topics where they lack expertise, especially once an AI speaks with the assured tone of an authority. That mental passivity becomes perilous during emergencies in remote terrain.

A compounding factor: chatbots are inclined to tell users what they wish to hear and present speculative advice as settled fact. Unlike social media posts, which merely inspire wanderlust, AI typically offers explicit instructions—right or wrong—alongside encouragement.

A Broader Pattern#

Poor outdoor planning fueled by technology is not new. A decade ago, social media destinations became deadlier as Instagram-hungry visitors sought photogenic, hazardous locations. Before that, smartphones and Google Maps lured people off-trail. Today's AI boom continues that trajectory.