Misuse Red Team - Research Engineer/Research Scientist
- Location
- Greater London, England, United Kingdom
society. The Misuse Red Team is a specialised subteam within AISI's wider Red Team. We red-team frontier AI safeguards for dangerous capabilities, research novel attack vectors, and develop advanced automated attack tooling. We share our findings with frontier AI companies (including Anthropic, OpenAI, DeepMind), key UK officials … other governments to inform their respective deployment, research, and policy decision-making. We have published on several topics, including novel automated attack algorithms (Boundary Point Jailbreaking), poisoning attacks, safeguards safety cases, defending finetuning APIs, third-party attacks on agents, agent misuse, and pre‐training data filtering. Some example impact ...