People + AI Guidebook

Jess Holbrook
Google Design
Published in
2 min readJun 24, 2019

The Google PAIR team is very proud to announce the launch of the People + AI Guidebook!

https://pair.withgoogle.com

This Guidebook will help you build human-centered AI products. It will enable you to avoid common mistakes, design excellent experiences, and focus on people as you build AI-driven applications.

It was written for UXers and product managers as a way to help bring a human-centered approach to AI product teams. That said, this Guidebook should be useful to anyone in any role wanting to build AI products in a more human-centered way.

There are six chapters, each with exercises, worksheets, and resources to help turn guidance into action:

User Needs + Defining Success

Data Collection + Evaluation

Mental Models

Explainability + Trust

Feedback + Control

Errors + Graceful Failure

Recommendations in the Guidebook are based on data and insights from dozens of Google product teams and a wide-ranging review of academic research. As our research continues to evolve, so will the way we approach AI product design. Expect updates to the Guidebook as the field of AI moves forward — think of it as a living document.

If you’re interested in a walk through of three of the chapters, here is a talk Kristie Fisher and I gave at our annual I/O developer event in May announcing the Guidebook:

You can also skip to a specific topic if you don’t want to watch the whole thing:

Intro | The role you play in creating human-centered AI products

  • 01:33 | Announcing the People + AI Guidebook

User Needs + Defining Success

  • 04:34 | Identifying if AI adds unique value
  • 08:43 | Optimizing for a clear, people-centered definition of success
  • 12:37 | Could a non-AI solution work?
  • 13:27 | Automation vs. augmentation

Data Collection + Evaluation

  • 15:47 | Translating user needs into data
  • 20:44 | Sourcing training data
  • 23:55 | Model tuning

Explainability + Trust

  • 25:48 | Calibrating trust
  • 33:29 | Displaying confidence

Recap | Takeaways and next steps

If you use the Guidebook, we would greatly appreciate you sending us any feedback you have.

Thank you to everyone who’s been supportive of our human-centered AI/ML efforts, especially the early ones. We think the Guidebook is a substantial step forward in the process…but far from the last one.

--

--

Jess Holbrook
Google Design

Head of User Research for Responsible AI at Meta. Formerly co-lead of the People + AI Research (PAIR) team at Google. Board Member, Forging Youth Resilience.