How a small Israeli startup was linked to rogue AI hacks at OpenAI, Anthropic and Meta
Rogue AI attacks traced to single source: Israeli startup Irregular linked to breaches at OpenAI, Anthropic, and Meta.
Why it matters
A coordinated campaign of AI system compromises across three frontier labs signals a new attack surface — autonomous agents — and raises urgent questions about agent security, supply-chain vulnerability, and the readiness of AI companies to defend against sophisticated adversaries.
The key facts
5 to knowIrregular (Israeli startup) linked to rogue AI attacks at OpenAI, Anthropic, and Meta
Multi-lab breach suggests coordinated campaign, not isolated incidents
Attack vector appears to involve agent systems or agent-adjacent infrastructure
Published August 9, 2026 — breaking security news with immediate practitioner relevance
Raises supply-chain and third-party risk concerns for AI deployments
Go to the source
CNBC Technologycnbc.com
Publisher excerpt: The rogue AI attacks involving OpenAI, Anthropic and Meta all tied back to a small Israeli startup named Irregular.