Human-Centered AI Design Principles

Most teams shipping AI features are making the same design decisions over and over without realizing it — how much the AI should act automatically, what to surface to users, where helpful ends and presumptuous begins. Without shared principles, every feature is a debate from scratch. This template gives you 10 pre-populated AI design principles across five categories — Human Control, Transparency, Personalization, Ethics, and Business Alignment — plus a Feature Review Log to make them operational and a Revision History to track how your thinking evolves. Run the task with your team to review, adapt, and ratify the principles in about 90 minutes. Then audit every AI feature against the principles your team has actually agreed on. Built for product designers and AI PMs at companies shipping LLM-based features. Grounded in a framework developed across Amazon Alexa, Google, and Uber by a Stanford AI design leader.