SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Anthropic is seeking a Safeguards Enforcement Analyst focused on Conventional Weapons to detect and mitigate misuse of its AI systems for real-world harm involving conventional weapons and dangerous technology. This role bridges policy, engineering, and threat intelligence to protect against weaponization of AI capabilities.
Key responsibilities include designing and architecting automated enforcement systems and review workflows that scale while maintaining accuracy; developing and maintaining evals that measure model performance on policy areas and surface regressions; partnering with Engineering and Data Science teams to optimize detection systems; reviewing flagged content to drive enforcement decisions and identify policy gaps; supporting the Safeguards policy design team with structured feedback on enforcement ambiguities; developing enforcement guidelines and reviewer documentation; staying current with emerging weapons trends, regulatory changes, and AI policy enforcement best practices; and identifying and escalating emerging misuse patterns, novel attack vectors, and coordinated violent activity.
The role requires deep applied expertise in weapons systems and the ability to translate complex technical evidence into enforcement decisions. Candidates should have experience in policy enforcement, threat intelligence, counterterrorism, government, or related fields with direct exposure to harmful content or dangerous technology. SQL and data analysis proficiency is essential, as is experience identifying emerging risks and communicating findings to diverse stakeholders (Product, Policy, Engineering, Legal). Experience with generative AI products, content moderation at scale, and understanding product policy implementation challenges is required.
Preferred qualifications include subject matter expertise in conventional weapons, autonomous systems, or critical infrastructure; familiarity with relevant legal and regulatory frameworks; experience developing evals or red-teaming AI systems; threat actor profiling and MITRE ATT&CK frameworks; tracking threat actors across surface, deep, and dark web; understanding of LLM capabilities and harm potential; Python proficiency; law enforcement or national security background; ability to assess technical plausibility and real-world harm potential; and OSINT and cross-platform threat analysis experience.
Note: This position involves exposure to explicit content spanning violent, graphic, hateful, or psychologically disturbing topics.
