Anthropic Product Policy Manager for Product Risk develops and executes risk assessment frameworks for AI product launches. Collaborates cross-functionally to drive safety mitigations and craft balanced policies.
Responsibilities
Develop and maintain risk assessment frameworks to identify and evaluate potential safety risks associated with new product features and functionality
Conduct and drive comprehensive product safety reviews, covering technical and non technical harms, to inform product launch and safety mitigation strategies
Collaborate closely with a variety of stakeholders including product and engineering teams and the broader Safeguards team to leverage deep policy, enforcement, and engineering expertise
Analyze the potential for misuse, unintended consequences, and harmful outputs of new model and product capabilities and translate this analysis into mitigations delivery.
Leverage SME risk assessments to inform overall product safety recommendations
Design and run bespoke evaluations for products that require tailored assessment and contribute to our product safety evaluations strategy.
Craft policy recommendations that strike a balance between enabling innovation and ensuring responsible AI deployment
Stay current on industry trends and emerging risks in AI development to proactively address potential issues
Contribute to regular reports on product policy risks and mitigations for senior leadership
Minimum Requirements:
Have a strong technical background while also feeling equally comfortable explaining highly technical concepts to non-technical stakeholders
Understand safety policies and safety considerations associated with a wide range of product surfaces
Qualification
Required
Strong technical background with ability to explain technical concepts to non-technical stakeholders
Understanding of safety policies and considerations across product surfaces
Adept at prioritizing safety mitigations in high volume launch processes
Experience conducting risk evaluations of novel products in fast moving organizations
