
Apollo Research
ACTIVEAn AI-safety research organization specializing in the study of "scheming" in frontier AI systems and in evaluating models and agents for deceptive, misaligned behavior.
Apollo Research
Founded 2023 · London, United KingdomApollo Research is an AI-safety research organization focused on "scheming" — situations in which advanced AI systems covertly pursue objectives misaligned with their developers' intentions. Its stated goal is to "secure frontier AI systems across development, deployment, and governance." Apollo builds a science of scheming, runs evaluations of models and agents, monitors agentic systems, advises governments, and formulates policy recommendations on loss-of-control AI risk. Its work includes partnering with OpenAI to test the o1 model before public deployment, publishing early evidence that frontier models can scheme in context, collaborating with OpenAI on anti-scheming interventions, analyzing internal AI deployment risks, producing a loss-of-control playbook for governments, and launching Watcher, an AI-agent monitoring tool. The organization was founded in 2023; most of its team works from London or San Francisco (the San Francisco office opened in 2026). In 2026, Apollo Research became a Public Benefit Corporation. Its CEO and co-founder is Dr. Marius Hobbhahn; its leadership also includes Chris Akin (COO), Dr. Charlotte Stix (Chief Governance Officer) and Dr. Alexander Meinke (Head of Research).
Founders
Co-founder and CEO of Apollo Research; an AI-safety researcher leading the organization's work on detecting and mitigating deceptive ("scheming") behavior in frontier models. He testified before the U.S. Senate on loss-of-control AI risks.
Domain of activity
1 areas · 3 technologies · 1 industries- AI Governance & Safety
- AI Security
- AI Governance
- Agentic AI
- Research & science
Global presence
HQ London · 2 offices · 2 regionsScale & funding
founded 2023Profile & metadata
4 classification · 3 external linksClassification
- Statusactive
- Development stagegrowth
- Listed onNo
- Organization typeAI Lab
External links