Novice Mount Shasta hikers rescued after relying on Google Gemini for planning
Three California hikers had to be airlifted from Mud Creek Canyon after the AI chatbot hallucinated packing requirements and estimated an impossible summit time.
Original cover art, generated for this story. THE VISSION does not republish third-party press imagery.
- Three novice hikers from Roseville, California, had to be rescued from Mount Shasta after using Google Gemini to plan their summit route and packing list.
- The AI model estimated an 8-hour round trip for a highly strenuous climb and advised the group to bring far less food and water than safety standards require.
- Authorities issued a public warning against using consumer AI chatbots for wilderness navigation, citing severe risks of geographic hallucination.
Three novice hikers from Roseville, California, had to be rescued from Mount Shasta after a series of critical trip-planning errors caused by their reliance on Google's Gemini AI chatbot. According to reports from the Siskiyou County Sheriff’s Office, the group utilized the large language model to design their summit itinerary and calculate their packing requirements, leading to an ordeal that highlights the real-world dangers of AI-generated safety advice.
The hikers reported that Gemini advised them to bring significantly less food and water than required and estimated a brief eight-hour round trip for a summit climb that typically demands intensive preparation and far longer transit times for novices. Following the AI's schedule, the group summited at 7:00 PM—seven hours past the recommended safety turnaround window. Becoming disoriented in the dark during their descent, they wandered into the treacherous Mud Creek Canyon, where one hiker suffered a severe knee injury, forcing the group to bivouac overnight.
Following a satellite SOS alert, the group was rescued on Monday morning by a joint team from the Siskiyou County Sheriff’s Office and U.S. Forest Service Climbing Rangers. In a public statement, search and rescue officials issued an urgent warning against using consumer AI chatbots for wilderness navigation and outdoor safety planning. Officials emphasized that chatbots regularly hallucinate crucial topographic details and survival guidelines, and urged the public to cross-reference all outdoor plans with official topographic maps and local ranger station reports.
This incident highlights the high-consequence risks of AI 'hallucinations' in everyday life. Because consumer chatbots present information with a highly confident and professional tone, users frequently mistake them for expert systems. As AI becomes deeply integrated into mobile operating systems, the boundary between general inspiration and safety-critical advice is blurring, with potentially life-threatening results in extreme environments.
Will Google implement proactive geographic safety filters in Gemini to block the generation of specific mountaineering and wilderness survival itineraries?
Still open. When the paper finds out, it will say so here and on the open questions page — including if it got this wrong.