Google's Gemini literally stranded hikers on a dangerous mountain — when you trust AI with your life, this is what can go wrong. The scariest part? The hikers specifically chose Gemini over other tools thinking it would be more reliable.
Three California hikers who used Google’s Gemini artificial intelligence assistant to plan their Mount Shasta climb had to be rescued after becoming stranded in a steep canyon, local authorities said.
The three hikers, all from Roseville, California, set out Sunday night to climb Mount Shasta, a 14,179-foot peak in Northern California, relying on a route generated by Google's Gemini. The AI directed them into a steep, off-trail canyon near the 8,000-foot mark, where one suffered a leg injury and the group could not self-rescue. A California Highway Patrol helicopter hoisted them out around 4 a.m. Monday after a ground team failed to reach them by foot. The Siskiyou County Sheriff's Office said the men were lucky to survive the overnight exposure at altitude.
Every major AI assistant, including Gemini, ChatGPT, and Claude, ships with hallucination rates high enough to produce fabricated trail names, wrong elevations, and nonexistent routes. For anyone using a chatbot to plan outdoor activities, the practical cost is now measurable: this rescue could have easily turned into a body recovery, and the hikers likely faced a four-figure bill if the helicopter wasn't a public asset.
This is the first widely reported case of a generative AI tool directly causing a wilderness rescue, but search-and-rescue teams have been flagging the trend for over a year as more hikers rely on apps like AllTrails, Gaia, and now LLMs for backcountry navigation. Traditional topographic maps and guidebooks require expertise; AI lowers the barrier to entry while quietly inheriting the same weaknesses that make crowdsourced GPS apps unreliable in remote terrain. Search-and-rescue callouts nationwide have roughly doubled since 2018, and the rise of AI trip-planning is now accelerating that curve faster than training programs can keep up.
Watch for the Siskiyou County Sheriff's Office to release a formal incident report within two weeks, which will clarify whether Google faces any liability questions or responds publicly. Google has not commented on whether it plans to add wilderness-specific safety guardrails to Gemini, but expect pressure from outdoor advocacy groups and possibly congressional scrutiny before summer hiking season peaks in July. The counterintuitive reality: the hikers chose Gemini precisely because they trusted Google's brand reputation, not realizing the model has no concept of whether a mountain actually exists.
When AI lies to the police: Anthropic’s model gave a fake murder tip, showing how rogue AI can cause real-world harm
Senate showdown: How AI policy became the battleground in the El‑Sayed vs. Rogers debate
Why most Americans think AI is moving too fast – and what it means for the future of AI regulation
How everyday Americans are letting ChatGPT help fill out their ballots—and why experts warn this could secretly skew election outcomes.
How a Hollywood movie featuring Sam Altman is sparking a nationwide debate about the real dangers of AI
Trump’s new AI task force: What government oversight could mean for the next wave of AI models