The Shape of AI

The Shape of AI is a resource that catalogs UX patterns for designing AI products, helping designers and builders create clearer, more transparent, and more trustworthy AI interactions across areas like guidance, inputs, tuning controls, safety “governors,” and trust indicators.

The Shape of AI’s philosophy is that AI products should be human-centered systems that solve real needs while staying legible, controllable, and trustworthy to the people using them. It treats AI UX as an ongoing two-way conversation, emphasizing clear inputs, visible reasoning, and adjustable controls so users can understand what the AI is doing, guide it, and recover from errors instead of feeling overwhelmed or sidelined. Underneath the specific patterns, it promotes principles like purposeful use of AI rather than “AI for its own sake,” logical transparency about how outputs are produced, tunability and personalization, strong feedback loops, and ethical integrity around privacy, bias, and honest identification of AI-generated content.