Human-Centered AI with Di Dang

Mark Mirchandani and Michelle Casbon take over the show this week to discuss AI and the PAIR Guidebook to Human-Centered AI. Mark Mandel pops in on the interview, and Di Dang, Design Advocate at Google, talks about her role in designing and building the guidebook with the intent of helping others create quality AI projects.

Di describes human-centered AI as a practice of not only being conscious of the project being built, but also considering how this AI project will impact us as humans at the end of the day. We influence machine learning so much, both intentionally and unintentionally, and it's our job to look at the project and results as a whole.

In the guidebook, topics like data bias in machine learning, what design patterns work, how to establish trust with the user, and more are addressed. Di explains that the guidebook is a work in progress that will develop with input from users and advances in technology.

Di Dang

Di Dang recently joined Google's Design Relations team as a Design Advocate supporting emerging technologies such as augmented reality and machine learning. Previously, she worked as a Senior UX Designer and led the Emerging Tech group at Seattle-based digital agency POP, advising clients on how VR/AR, web/mobile, conversational UI, and machine learning could benefit their end users. With a degree in Philosophy and Religion, she considers herself an optimistic realist who is passionate about ethical design. You can find Di onstage doing improv or on Twitter @dqpdang.

Cool things of the week

• Bringing the best of open source to Google Cloud customers blog
• James Ward's Cloud Run button site
• Michelle's favorite codelabs from I/O
  • TPU-speed data pipelines site
  • Your first Keras model site
  • Convolutional neural networks site
  • Modern convnets, squeezenet, with Keras with TPUs site

Interview

• People + AI Guidebook site
• PAIR site
• GCP Podcast Episode 114: Machine Learning Bias and Fairness with Timnit Gebru and Margaret Mitchell podcast
• Machine Learning Crash Course site
• Google Clips site
• Google Brain Team site

Question of the week

• How do I get started with practical AI?
• Build an Appointment Scheduler Chatbot with Dialogflow

Where can you find us next?

Michelle will be at Google I/O and Kubecon Europe.

• No I/O event in your area? You can host one!