Director, AI Security

Gibson Dunn is a leading global law firm, advising clients on significant transactions and disputes. Our exceptional teams craft and deploy creative legal strategies that are meticulously tailored to every matter, however complex or high-stakes. The firm’s work is distinguished by a unique combination of precision and vision.

Based in New York, the Director of AI Security will work directly with the CISO to define the firm’s AI security strategy and determine and build the function required to deliver it — its shape, its size, its sequencing and its budget. This is the firm's first dedicated AI security role, with no inherited team or playbook; defining the team required to deliver the strategy is a key part of the role. The position requires the ability to engage credibly with senior executives and practice group leadership on strategy, while also reasoning about technical details such as token scopes and prompt injection.

This role reports to the Chief Information Security Officer.

Responsibilities include:

AI Security Strategy & Program Development

• Define the firm's AI security strategy, target architecture, and multi-year roadmap in partnership with the CISO, and secure executive endorsement.
• Design, budget, recruit, and lead the AI security function, including team structure, tooling, and build-versus-buy decisions.
• Establish the security control framework for AI systems.
• Own the security review stage of the AI use-case intake and approval process, ensuring a clear and timely path from proposal to production.
• Partner with the Office of General Counsel and the Cyber & Data Governance Committee on obligations relating to confidentiality, privilege, and competent use of technology.

 Stakeholder Engagement & Secure AI Enablement

• Build relationships with partners, practice group leaders, and business leaders to understand the drivers of AI adoption and translate business needs into secure delivery options.
• Present AI risk and strategy to executive stakeholders and, where required, to clients.
• Conduct security architecture and design reviews for AI platforms, RAG pipelines, agent frameworks, and in-house builds.
• Define controls against AI-specific threats, including prompt injection, tool abuse, data leakage, supply chain compromise, and cross-matter contamination.
• Address unsanctioned AI use through discovery, sanctioned alternatives, and clear guidance, in partnership with IT and the practice groups.

 Identity & Access Management for AI Systems