EXPLORINEER
en

VN — VnExpress

The Risk of Getting Trapped When Following AI Directions

AI2 min read

Hikers and tourists are relying on AI chatbots to plan outdoor trips, and some are getting into serious trouble. Rescue teams in the US and Canada say the bots can give confident advice that is wrong or unsafe.

What happened

More people are using AI chatbots to plan trips into wild areas. Some of these tools mix unreliable information from social media into maps and route advice that look trustworthy but are not realistic. Several recent rescues show how this can go wrong.

Key facts

  • Late August: three climbers on Mount Shasta in Northern California needed rescue after they climbed too late, did not bring enough food, and overrated their fitness. They blamed part of their poor planning on Google's Gemini chatbot.
  • Early October: a 16-year-old boy was stuck on a rock ledge in British Columbia, Canada. A helicopter was needed. He had used Anthropic's Claude chatbot to plan the trip.
  • A survey by Phocuswright found about 58% of US travelers use AI to plan routes. Over 60% of Millennial and Gen Z travelers do.
  • One study found a travel planning chatbot was correct only 10% of the time. That study was done in 2024, and the tools have improved since then.
  • At Grand Teton National Park in Wyoming, incidents linked to visitors using AI for travel info rose from 0 to 26% over the past year.

Why chatbots make mistakes

Chatbots answer with a lot of confidence, even when their information is wrong. They rely on fixed data that may be old or incomplete. They also tend to say what users want to hear. A researcher warns that people who stop thinking critically may find it hard to think clearly in a scary situation far from help.

Why it matters

Mistakes in the wilderness can be deadly. A bad route, a wrong map, or a false claim that a campsite is allowed can put people in danger and tie up rescue teams. Experts say it is still unclear whether AI misuse is causing more accidents, so people should check AI advice against official maps, park rules, and weather information.

What to watch

  • Whether rescue teams see more calls linked to AI advice over the coming years.
  • Whether chatbot makers make their tools more accurate for outdoor planning.
  • Whether national parks add clearer warnings about relying on AI for trip planning.

Glossary

Gemini
Google's AI chatbot that can answer questions and make plans.
Claude
An AI chatbot made by the company Anthropic.
Phocuswright
A research company that studies the travel industry.
Grand Teton National Park
A large protected nature area in Wyoming, USA, where people hike and camp.

Ad

Add 150 words. Learn 10 a day.

Dicter word books send only the words you learn to your review notes.

More in this topic

AITechCrunch

After blocking Muse, Amazon now lets users text Alexa to help them shop

Amazon has launched a feature that lets select U.S. customers chat with its Alexa AI agent over text messages to search for items, add them to cart, and make purchases. The rollout comes months after Amazon blocked Meta's Muse from querying its site, and the company plans to expand access in the coming weeks ahead of the holiday shopping season.

AITechCrunch

Underdog shows off a 32MB speech model in new demo

Underdog has released a browser-based game built with its Woof voice model, which is only 32MB and has 31 million parameters. The company says the model runs on most devices without an internet connection, and it is open for anyone to download. The demo follows Underdog's Bark speech-to-text model, a 167 MB system for transcription and dictation designed for edge devices.

AIThe Register

Bank of England pours cold water on AI productivity boom

The Bank of England's latest Monetary Policy Report estimates AI has raised UK productivity by about 0.5 percent over the past three years, or 0.1 to 0.2 percentage points of additional annual growth since early 2024. The bank said much of AI's contribution to GDP growth comes from developing and deploying the technology rather than from businesses using it to become more productive. Other surveys cited in the article also report limited returns on AI spending so far.

AITechNode

Apple’s China AI partner has had a rough ride

Apple Intelligence began beta testing in China on Nov. 4, with Alibaba's Qwen available in selected Siri, Writing Tools and Visual Intelligence features. Qwen's rise was not smooth: it faced internal skepticism, fierce competition from DeepSeek, and senior departures, including technical leader Junyang Lin in March 2026. Alibaba responded by consolidating teams under Token Foundry in June and naming Dayiheng Liu head of the Qwen LLM project in September.