Senior AI Policy & Cyber Safeguards Lead

Society for AI Security; San Francisco (CA); Hybrid; USD 180,000 - 280,000; Full time

Job summary

Anthropic in San Francisco, CA is seeking a senior policy leader to govern access to its sensitive cyber-relevant model capabilities. You will consolidate our controlled-access framework, update policy commitments, and serve as the primary policy contact with government and regulators.

The role requires deep engagement with regulators, partners, and the coordination around vulnerability disclosure. You won't write code, but you will translate complex technical materials—probes, classifiers, and

Qualifications

• Bachelor's degree or equivalent required.
• Experience building and defending policy positions.
• Strong ability to translate technical material into policy.
• Experience engaging with government and regulators on cybersecurity or AI matters.

Responsibilities

• Own access-requirements policy and consolidate controlled-access framework.
• Keep cyber-related policy commitments current with regulatory standards.
• Serve as primary policy interface with government and regulators.
• Build relationships across vulnerability disclosure, cybersecurity policy, and national security communities.
• Author and maintain reporting to partners and regulators.
• Translate technical evaluations into policy positions for external engagement.
• Engage with technical teams to understand capabilities relevant to access policy.
• Demonstrate experience engaging with US government and regulatory parties on cybersecurity.

Skills

Policy leadership

Regulatory engagement

Government liaison

Technical policy reading

Education

Bachelor's degree

Job description

Anthropic in San Francisco, CA is seeking a senior policy leader to govern access to its sensitive cyber-relevant model capabilities. You will consolidate our controlled-access framework, update policy commitments, and serve as the primary policy contact with government and regulators.

The role requires deep engagement with regulators, partners, and the coordination around vulnerability disclosure. You won't write code, but you will translate complex technical materials—probes, classifiers, and