AI chatbots linked to dangerous backcountry hiking errors

Park officials and rescue teams are reporting an increase in backcountry mishaps attributed to hikers relying on artificial intelligence for trip planning. On Mount Shasta in Northern California, three hikers required rescue after making poor decisions regarding food, fitness, and timing, which they admitted was partly due to following advice from Google’s Gemini chatbot.

Other incidents include a visitor at Mount Shasta attempting to camp in a prohibited area based on AI information, and a hiker using an AI-generated map that incorrectly placed mountain features. In British Columbia, a 16-year-old hiker had to be rescued from a ledge on Crown Mountain after using Anthropic’s Claude chatbot to plan his excursion.

In Grand Teton National Park, Wyoming, officials have noted a significant rise in visitors using AI for travel information, with the reported usage increasing from zero to 26 percent over the past year.