Threat Modeling Lead, Harmful Manipulation, DeepMind
New York, NY, USA +2 · Full-time · Posted Oct 14, 2026
- Location
- New York, NY, USA
London, UK
Mountain View, CA, USA - Education
- PhD degree in Computer Science, Political Science, Psychology, Computational Social Science, a related field, or equivalent industry experience.
About the job
Develop threat models and evaluations of harmful manipulation risks in advanced AI models to inform safety thresholds, deployment safeguards, and launch decisions.
What you'll do
- Maintain harmful manipulation threat models and integrate them into Frontier Safety Framework risk assessments.
- Design and implement evaluations of harmful manipulation capabilities and risks; publish novel methods and findings.
- Assess deployment safeguards and engage with external researchers, stakeholders, and experts on frontier AI risks.
What they ask for
- PhD in Computer Science, Political Science, Psychology, Computational Social Science, or a related field, or equivalent industry experience.
- Experience performing technical work such as OSINT data collection, API scraping, log analysis, or other data analysis.
- Experience studying, tracking, or evaluating human manipulation, persuasion, propaganda, influence operations, social engineering, or related social phenomena.
Nice to have
- Experience in a national laboratory, government organization, military intelligence unit, specialized corporate/frontier lab threat intelligence team, or specialized research institution.
- Experience communicating complex intelligence and technical risks as actionable insights for business leaders, governance bodies, and policy experts.
Pay
$207000 - $300000 USD base salary
Tools
OSINT, API
AI focus
Evaluate how advanced AI models interact with and influence users, identify harmful manipulation risks, and inform safety evaluations and deployment safeguards.
Model-written summary · the original posting is authoritative