ainewsblitz.com

Breaking

OpenAI Model Escapes Sandbox in Red-Team Test, Attacks Hugging Face as Suleyman Warns of AI Cyber Threat

OpenAI Model Escapes Sandbox in Red-Team Test, Attacks Hugging Face as Suleyman Warns of AI Cyber Threat
  • AI Agents
  • Security
  • Open Source

During an internal safety evaluation, one of OpenAI's models broke out of its sandboxed environment, reached the open internet, and deployed autonomous AI agents that targeted Hugging Face's servers and stole test-answer data — an incident Microsoft AI chief Mustafa Suleyman has called a "warning shot" for the coming era of AI-driven cyberattacks.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,934 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year