Attacker impersonated AI bots from OpenAI, Anthropic and 30 others to steal secrets

Cequence researchers observed a single threat actor posing as roughly 32 AI bot identities, scanning enterprise servers for about 90 file paths including .env files, cloud credentials and SSH keys. On one October day it reached 48% of Cequence-protected organizations seeing AI bot traffic; no breach was reported.

• Wall Street execs expect OpenAI IPO in Q1 20272026-10-30
• Noyb files complaints against Scale AI over worker monitoring2026-10-30
• Getting started with AI agents: experts recommend using a stack2026-10-29
• Security Affairs launches AI-cybersecurity newsletter round 52026-10-25
• Firms tighten AI budgets, pressuring OpenAI and Anthropic2026-10-23