This collection of articles and interviews surveys human-centered approaches to machine learning that can make AI more human-friendly, usable, and ethical. It provides a handbook for students, researchers, and practitioners who want new ways of approaching AI that place humanity at their center. It shows how to apply methods from human-computer interaction to the new technologies of AI and ML with a view to enabling computing technology to become user-friendly and human-centric. The book has 13 articles and 9 interviews from a range of different perspectives, helping readers understand existing machine learning systems and their impacts on people and society. It is an ideal introduction both for human-computer interaction practitioners who are interested in working with ML and for ML experts interested in making their practice more human-centered. The book offers a critical lens on existing machine learning alongside an optimistic vision of AI in the service of humanity.

‘Human-Centered Machine Learning grounds us in the perspectives and techniques we need to shape AI as a benefit for society by being explicit about what we already know about designing technology for people and humanity and how that applies to machine learning. A book to ground us in this moment of rapid change.’

Cecily Morrison - Microsoft Research

‘Machine learning is already human-centered-just not always on purpose. Built from our data, shaped by our choices, and deployed into our messy world, it reflects us. This collection shows how to make that explicit: with clear principles, real practices, and better judgment about what we build and why.’

Jess Holbrook - Head of UX Research for Microsoft AI