Latest Irregular news, funding rounds and robotics & AI coverage. 2 articles tracked by The Robotics Media.
Irregular is an AI security firm that runs red-team and blue-team programs against frontier language models and agent systems for labs including OpenAI, Anthropic and Meta. The company has published high-profile evaluations of model misuse, agentic side-effects and model-training attacks, and closed an $80 million Series A in mid-2026.