The past decade has witnessed an unprecedented acceleration in both the sophistication and uptake of various algorithmic decision tools. From music and TV show recommendations, product advertising, and opinion polling to medical diagnostics, university admissions, job placement, and financial services, the range of the potential application of these technologies is truly vast. And while the business sector has wasted no time getting on board, governments too have been steadily integrating algorithmic decision support systems into their daily operations. Many police and law enforcement agencies around the world, for example, have co-opted deep learning tools in an effort to optimize efficiency and (so they say) reduce human bias. But while the roll-out continues to gather momentum and enthusiasts have welcomed the dawn of a new era, not everyone is convinced. Must those awaiting the outcome of a health insurance claim or defendants seeking bail or parole simply take it on faith that a machine knows best? Can a machine deciding such matters as the likelihood of a criminal reoffending really be accurate, free from bias, and transparent in its operations?
This book is the result of a joint effort in trying to make sense of the new algorithmic world order. It’s structured around ten core themes. These canvass such questions as, what is artificial intelligence? Can an AI explain its decisions? Can it be held legally responsible? Does it have agency? What kind of control should humans retain over such systems, and does it depend on the kind of decision being made? Is the law of privacy in need of a fundamental rethink now that data sharing is easier and far more common than even ten years ago? How do we address the potential for manipulation through targeted political advertising? Is the use of decision tools by governments interestingly different from their use in industry? Do states have unique obligations to their citizens in this regard? How best to regulate behemoths like Facebook, Google, and Apple? Is regulation the answer? What kind of regulation?
The authorship of the book is a little unusual and perhaps merits a word or two of explanation. Put simply, although we wanted to have the best people write on the topics of their expertise, we didn’t want the book to be a clunky miscellany of chapters and jarring styles, as is all too often the case with edited collections. Producing an authoritative work for the general public was certainly important to us, but we were adamant that it should have a single arc and speak with one voice. So, we needed someone to write a good chunk of the book—to set the tone, if you will—and someone who’d then be happy to take on what the others had written and mold it to that style. Ideally these roles would fall to the same person, and John Zerilli agreed to be that person, contributing just under half of the material and reshaping the rest to jibe with what he’d written.
The result, we trust, will prove useful to the bemused citizen.