How bad data keeps us from good AI

If you think today’s tech giants invented data mining and market manipulation, then think again. Jill Lepore’s eye-opening historical account reveals how brilliant social scientists and Madison Avenue ad firms used computing power in the 1960s to develop an early version of a data machine to simulate and predict voting and consumer behavior. Lepore claims that, among other successes, the machine successfully supported the campaigns of Eisenhower and Kennedy. These data scientists built a very early version of today’s digital and social platforms that exploit consumer attention, commodify information and undermine democracy. While Simulmatics — the corporation — failed, the foundational science remains core to all digital businesses. If Then is a cautionary tale.

A Turing Award-winning computer scientist and statistician shows how understanding causality (the study of cause and effect) revolutionized science and will revolutionize artificial intelligence. This book examines the notion of "correlation is not causation," cutting through the history and confusion around the second piece: causality. Pearl claims that most AI algorithms describe associations and patterns without answering the "why" question. In addition to feeding algorithms relevant data, Pearl shows how, by incorporating causality and counterfactuals, we can can build purposeful AI and intelligent machine for the humanity.

This profoundly ambitious and original book answers the dystopian questions: What happens when machines surpass humans in general intelligence? Will artificial agents save or destroy us? Bostrom discusses the possibility of machines taking over humans, but notes that we have one advantage: we get to make the first move, set the initial conditions and design to mitigate adverse outcomes. This is an important and timely book for all interested in the future of AI and its impact in our lives.

This book — which covers the social history of the American census from its origins to present — is essential reading for all those who are interested in the census, historical or current. The recent edition has been updated to trace census developments since 1980, including the undercount controversies, the arrival of the American Community Survey and innovations of the digital age. Anderson’s scholarly text effectively bridges the fields of history and public policy, demonstrating how the census both reflects the country’s extraordinary demographic character and constitutes an influential tool for policy making.

A simple and practical guide to building ethical AI organization. This guide covers dos and don'ts, including how to operationalize data and AI ethics, as well as building an ethical data infrastructure.

This podcast is a terrific resource towards understanding how AI is impacting our lives, industry, government and society. Azeem Azhar, curator of the Exponential View newsletter, explores these questions in deep conversations with the world's leading AI thinkers and practitioners.