Oxford Witt Lab

Website of the Oxford Witt Lab (AI research group)

Oxford Witt Lab Search Oxford Witt Lab for Trust in AI The Oxford Witt Lab (OWL) pioneers foundational assurance for AI-human-institution ecosystems. From our work on undetectable threats and multi-agent security to principled evaluations and governance interventions, OWL builds the theoretical and empirical backbone for trust in advanced AI. We work independently, openly, and rigorously to advance anticipatory security in complex sociotechnical systems. We offer an assurance clinic that advises on mitigating risks in AI deployments of any kind - whether in industry or academia. Email [email&#

saved by

• N . Soma Sekhar

related reading

• Student Projects / Supervision | Oxford Witt Labwittlab.ai
• [2607.07368] Multi-Agent AI Control: Distributed Attacks Hamper Per-Instance Monitorsarxiv.org
• AI Risk & Reliability - MLCommonsmlcommons.org
• Research Mission | Oxford Witt Labwittlab.ai
• Patterns and problems in multiagent systemsanthropic.com
• Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com
• Red-teaming a network of agents: Understanding what breaks when AI agents interact at scale - Microsoft Researchmicrosoft.com
• Of Swarms and Sand Godsblog.cosmos-institute.org
• Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
• Lessons from Moltbook and OpenClaw: The Agentic Internet’s Trust Problem - Irregularirregular.com
• How we built our multi-agent research system \ Anthropicanthropic.com
• The End-State Fallacy: Where Is AI Security Headed?endstatefallacy.com