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
Google DeepMind Blog · November 10, 2026

ASIMOV-Jailbreak: test robotic agent decisions and benign refusals

analysis Model EvaluationAgent SecurityAdversarial ML
Why it matters

DeepMind’s robotics release includes an inspectable ASIMOV-Jailbreak evaluation suite. Its dataset card describes paired harmful and benign tasks and multi-turn adversarial episodes, combining tool-call checks with an automated judge. This provides a method for evaluating agent decisions; model-reported benchmark resistance and factory demonstrations do not establish physical safety in a new deployment.

My takeaway: Run the harmful and benign suites separately, retain model and judge versions, and inspect failures and over-refusals. Evaluate multi-turn escalation as well as single requests. Keep physical interlocks and supervised hardware validation separate from these simulated decision tests.
Keep exploring

More curated notes connected through Model Evaluation and Agent Security.

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.