Back to Latest SignalTECHMEME / MODEL RELEASE

Vals: in simulated scenarios, AI models from Anthropic, Google, xAI, DeepSeek, and OpenAI identified civilians, ambulances, and a school as military targets (Katrina Manson/Bloomberg)

Source: TECHMEME · NOV 17, 2026

Katrina Manson / Bloomberg : Vals: in simulated scenarios, AI models from Anthropic, Google, xAI, DeepSeek, and OpenAI identified civilians, ambulances, and a school as military targets — Advanced artificial intelligence models in simulated scenarios regularly misidentified civilians, ambulances …

IN THIS STORY
  • AI Safety
  • OpenAI
  • Anthropic
  • Alphabet
  • DeepSeek
  • xAI

Sources

  • Read the original at TECHMEME

Related on Frontier

  • SEC filing: Amazon and Google are set to hold nearly 15% and 14% of Anthropic's outstanding capital stock, respectively, after its IPO and private placement (David Morris/Bloomberg) Nov 17, 2026
  • OpenAI Unveils New Astra AI Model After Scrapping Prior Release Nov 17, 2026
  • California Sues OpenAI, Seeking Court-Ordered Changes After the Hugging Face Hack Nov 17, 2026
  • Google spots the first hacker AI system that finds and exploits flaws autonomously Nov 17, 2026
  • Gemini AI agents repeatedly bypassed safeguards to attack real website Nov 16, 2026
MORE LATEST SIGNALSIGNALS