Urgent.News

What's breaking now, across thousands of outlets.

AI

Hikers 'relied too much' on AI and had to be rescued. A ranger says it wasn't the first AI-related mishap he's seen.

Three hikers who needed to be rescued from Mount Shasta after they "relied too much" on AI are part of a new trend, a US Forest Service ranger said.

Three hikers were rescued from Mount Shasta after relying too heavily on AI for their trip, according to the Forest Service. Sheriff's Office spokespersons reported that the climbers had used Google Gemini, YouTube, and the AllTrails app to plan their ascent. The group admitted to rangers that they had "relied too much on AI rather than their own critical thinking," Business Insider reported.

Ranger Nick Meyers, a lead climbing ranger for the US Forest Service, said this was the third AI-related mishap he had witnessed this year. One woman, according to Meyers, ignored a "no camping" sign because AI had told her it was permissible. Meyers, who is also the director of the Mount Shasta Avalanche Center, said that he has seen a growing trend of hikers making mistakes due to AI assistance in 2026.

The most serious incident involved three 20-year-old men who got lost after summiting Mount Shasta, California's fifth-highest peak, and had to be rescued. They had established a base camp at a prohibited site and, without adequate supplies, spent a night in a makeshift camp after their cellphone's navigation died. Meyers advised hikers to use AI as one tool in their planning process, but emphasized the importance of critical thinking and experience.

Written by urgent.news from Business Insider's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at businessinsider.com →

More in AI

Seattle Times, Newsday sue OpenAI and Microsoft; allege copyright infringement

The Seattle Times and Newsday sued OpenAI and Microsoft on Friday, alleging in federal court that the tech companies copied the newspapers’ journalism without permission to train their AI systems. The suit in the US District Court for the Southern District of New York alleges that OpenAI and Microsoft scraped the newspapers’ websites…

Why Most AI Agents Fail in Production

The demo is flawless. The agent reads the support ticket, categorizes it, looks up the customer, drafts a response, updates the CRM, and posts a summary in Slack. Everyone nods. Someone says, “This is going to save us hundreds of hours.” Two weeks later, the same agent has replied to the wrong customer, created thirty-one duplicate tasks…

  • AI agents fail in production due to insufficient engineering controls.
  • Common failure modes include overprivileged tools and inadequate evaluations.
  • To improve, separate agent's read/write access for better control.

More from Saturday 5 September →