Creating Great User Experiences with Imperfect AI

Machine learning is arguably the most important trend in technology today. From predicting what you’re going to say before you say it to fabricating lifelike videos, ML is a powerful technology.  

But designing great user experiences for ML-powered features presents unique challenges to a product designer: How do you create a mental model for a system with thousands of parameters? How do you explain recommendations from a system that has no rules? How do you provide user control? How do you respect user privacy expectations and tackle issues like ML fairness? 

In this talk, Annika Crowley (Interaction Designer @ Google) discusses the challenges she and her team encountered and techniques they developed in building machine learning features in Gmail and Inbox. She also shares strategies that they’ve learned over the last couple of years at Gmail and Inbox about how to create great user experiences with imperfect AI.  

Watch next, Paul Lambert, Product Manager at Google discussing how to build great machine learning experiences for Gmail.