AI Red Team Engineering notes, research references, and field notes on AI systems, responsible AI, governance, and compliance.
Global search

Search AI Red Team

Search across the complete knowledge base.

Enter at least two characters to begin.
↑↓ select Enter open Browse the full research archive
Proofpoint · October 1, 2026

TA419 used policy-expert impersonation to phish Microsoft 365 sessions

analysis AI Red Teaming
Why it matters

Proofpoint’s October 2026 investigation describes targeted phishing against US AI policy researchers. Benign invitations impersonating recognizable experts established a conversation before a file-sharing link led to a proxied Microsoft sign-in. A modified browser-in-the-browser kit relayed authentication and captured session cookies, including after supported MFA challenges. Proofpoint attributes the activity to the China-aligned group TA419; the report documents infrastructure and targeting, rather than a count of confirmed account compromises. The AI connection is the selected victims and policy pretexts, not evidence of an AI-driven attack.

My takeaway: Verify unexpected research invitations through an independent channel. For exposed accounts, investigate session activity as well as password theft, and prioritize origin-bound, phishing-resistant authentication.
Keep exploring

More curated notes connected through AI Red Teaming.

OpenAI News · framework

Pacing model development in an era of cyber-critical capabilities

OpenAI says preliminary evidence that Astra may meet its Critical cybersecurity threshold led it to pause frontier reinforcement-learning work for two weeks and keep its largest planned run on hold. New safeguards include stronger workload and network isolation, continuous boundary testing, token-level monitoring that escalates suspicious tool activity, and broader alignment checks for deception, reward hacking, and unauthorized access.

OWASP GenAI Security Project · guide

OWASP Top 10 for Agentic Applications for 2026

OWASP's community guide organizes agentic-system risk into ten categories, including goal hijacking, tool misuse, identity and privilege abuse, memory poisoning, insecure inter-agent communication, cascading failures, and rogue-agent behavior. It provides a shared taxonomy and mitigation starting point rather than a certification checklist or evidence that a deployed system is secure.

OpenAI News · framework

A blueprint for democratic governance of frontier AI

OpenAI proposes a three-part U.S. frontier-AI governance model: harmonize emerging state safety laws into a federal baseline, strengthen CAISI as an evaluation and standards institution, and coordinate a broader resilience program. Proposed controls include severe-risk evaluations, transparency reports, independent audits, safety-incident reporting, model-weight security, whistleblower protection, and periodic technical assessments.