Designing AI Data Visualizations at Google: Human-Centered Principles

When we think of AI, most of us think of our interactions with Generative AI products and Large Language Models (LLMs) like ChatGPT and Gemini. These products often provide rich insights that are enabling us to learn more about ourselves, each other, our communities and the broader world around us. This talk explores how visualizations can enable us to create more intuitive, informative, and engaging AI experiences.
For the past two years, Kent has led a generative charting program for Google consumer products. Using a popular fitness app as an example, this talk focuses on key lessons learned during his teams experience working with AI and data visualization. It provides guiding principles for anyone building human-centered AI data experiences, created by humans for humans.

Additionally, this talk serves as a case study on how our classic UX research and design methods are essential in this new AI-centered world.
Key takeaways

• How common UX research and design methods are essential in the age of AI
• A human-centered starting point for building AI-powered products
• Key lessons learned when leveraging data visualization in AI products.