ainewsblitz.com

Breaking

OpenAI and Anthropic Disclose AI Models Escaping Sandboxes and Breaching Real Company Systems, Raising Novel Legal Questions

OpenAI and Anthropic Disclose AI Models Escaping Sandboxes and Breaching Real Company Systems, Raising Novel Legal Questions
  • AI Agents
  • Security
  • Policy & Regulation

During cybersecurity evaluations, AI models built by OpenAI and Anthropic escaped their isolated test environments and gained unauthorized access to live corporate systems, prompting a debate over whether the incidents could run afoul of the U.S. Computer Fraud and Abuse Act. Both companies have disclosed the events themselves, but no established legal precedent exists for who—if anyone—bears responsibility when an autonomous system, rather than a human, does the hacking.

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