General AI
AI Chatbots Leading Hikers into Danger
WIRED AIhigh confidence
Why it matters
- →AI hallucinations are causing real-world physical harm in outdoor recreation.
- →User trust in AI is leading to 'cognitive surrender' and reduced critical thinking.
- →Park services are seeing rapid adoption of AI for planning, increasing safety risks.
Search and rescue personnel across the US West are reporting a rise in backcountry incidents linked to users relying on generative AI for trip planning. Grand Teton National Park data shows a sharp increase in visitors using AI for travel information, rising from zero to 26 percent over the past year. Experts warn that chatbots like Gemini and Claude often provide plausible but incorrect advice, leading to 'cognitive surrender' where users fail to verify facts with official sources. While overall rescue calls have fluctuated, officials emphasize that AI's authoritative tone can dangerously mislead inexperienced outdoor enthusiasts.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1Hugging Face Explores Boundary-Aware AI Safety
Hugging Face Blog · September 8, 2026 · Background
- 2AI Agents Hacking for Answers Raises Safety Alarm
MIT Technology Review AI · September 23, 2026 · Background
- 3OpenAI agents bypassed safety shutdowns via DNS
Wes Roth · September 26, 2026 · Background
- 4White House launches AI chatbot for government services
TechCrunch AI · September 29, 2026 · Background
- 5AI Hallucinations Disrupt Customer Service
The Verge AI · October 2, 2026 · Background
- 6The rise of SMS-native AI agents
TechCrunch AI · October 10, 2026 · Background
- 7AI in Emergency Alerts: Trust vs. Accuracy
WIRED AI · October 28, 2026 · Background
More from WIRED AI
© WIRED AIMarket & Regulationother
400+ AI Apps Hide Deepfake Networks
A coordinated network of over 400 apps masquerading as benign tools like pet portrait generators is secretly serving sexually abusive deepfakes. Researchers discovered the infrastructure after tracing explicit ads on Meta platforms to these deceptive applications, which use server-side logic to toggle between safe and harmful content based on user origin. The operation, linked to a Hong Kong entity and Chinese developers, bypassed App Store reviews by swapping approved code for malicious interfaces post-launch. This reveals a systemic failure in platform enforcement where disposable apps can be rapidly deployed and monetized before detection.
WIRED AI
© WIRED AIResearchother
AI-Augmented Infostealer 'Volta-Autodoxer' Targets High-Value Victims
Cisco Talos has uncovered a new infostealer strain that actively uses commercial LLMs to profile victims and prioritize high-value targets. Dubbed Volta-Autodoxer, the malware collects credentials and browsing data, then feeds this into models like Claude Opus 4.6 via OpenRouter to generate structured intelligence dossiers. This marks a shift from passive data theft to active, AI-driven triage, allowing attackers to bypass guardrails with fallback models like Kimi K2 or Gemini. The capability foreshadows a future where lower-sophistication criminals can execute sophisticated, targeted attacks at scale.
WIRED AI
© WIRED AIAgentsagents
AI Agent Performs Live Mouse Brain Injection
Exin Therapeutics has demonstrated an AI agent physically controlling a robotic arm to inject drugs into a live mouse's brain. The experiment reveals a critical divergence in model behavior: Anthropic’s Claude Opus refused the procedure due to safety concerns, while Sonnet proceeded. This suggests that as agents gain direct access to physical lab equipment, their alignment and refusal mechanisms become operational realities rather than theoretical safeguards. It marks a shift from AI merely analyzing data to actively manipulating biological systems under human supervision. The team initially tested Claude Opus 5.5, but it halted the injection, citing potential harm. Switching to Claude Sonnet 5 allowed the procedure to complete, though with delays caused by coordinate system glitches. A secondary Opus agent had to fix code errors without knowing a live surgery was underway. Human oversight remains essential as researchers test more complex tasks like craniotomies.
WIRED AI
More in General AI
© TechCrunch AIGeneral AIother
Google expands Dreambeans AI stories to India
Google’s experimental personal story app Dreambeans has left the U.S., launching in India with support for 13 local languages. The app pulls data from Gmail, Photos, and Calendar to generate illustrated narratives, recently adding Veo-powered animations. Despite high-profile competitors like Meta’s Muse, Dreambeans saw modest U.S. adoption, raising questions about its mass-market appeal. The expansion tests whether personalized AI memory tools can scale in a multilingual market where code-switching is the norm.
TechCrunch AI
© TechCrunch AIGeneral AIproductivity
OpenAI merges Chat and Work tabs in desktop app
OpenAI is collapsing the friction between casual chat and project work by merging the Chat and Work tabs into a single unified interface. This move eliminates the mental overhead of switching contexts, allowing users to seamlessly transition from brainstorming to executing tasks with files and plugins without leaving the main view. While Codex remains distinct for deep coding, the token pool consolidation means usage is now tracked holistically across both modes. It’s a necessary UI cleanup that finally aligns ChatGPT’s desktop experience with how people actually use it.
TechCrunch AI
© MIT Technology Review AIGeneral AIagents
The Agentic Web: Agents Replace Browsers
Tim Berners-Lee’s Semantic Web dream is finally arriving, but not through the structured tags he envisioned. Instead, a new 'agentic web' is emerging where LLM-based agents navigate or interface directly with services via connectors like Stripe and Gmail. This shift moves the internet from human-clicked pages to machine-readable transactions, fundamentally altering how we interact with digital platforms. Major sites are already optimizing for this by serving content specifically to AI crawlers rather than human browsers.
MIT Technology Review AI