AI Principles reviews and operations

Any team can request AI Principles advice. Reviewers also consider an ongoing pipeline of new AI research papers, product ideas, and other projects. Reviewers identify relevant AI Principles as frameworks for action.

Reviewers analyze the scale and scope of a technology’s potential benefits and harms. Similar to a case law process, they apply precedents. Reviewers consult with internal experts on privacy, security, ML fairness, and other domains as needed.

Reviewers recommend technical evaluations (e.g., checking for unfair bias in ML models). If necessary, they consult with trusted external advisors (e.g., human rights experts). Reviewers offer mitigation strategies; adjustments are made.

Reviewers decide whether to pursue or not pursue the AI application under review (e.g., Cloud AI Hub and text-to-speech). If multiple products might be affected, a council of senior Google executives decides. The decision can become a new precedent.