“Algorithms have made our lives more efficient, more entertaining, and, sometimes, better informed. At the same time, complex algorithms are increasingly violating the basic rights of individual citizens.”


The world is currently in a maelstrom of the once-in-a-generation pandemic now made even more chaotic by ongoing international protests against racism. The exceedingly timely arrival of The Ethical Algorithm is one hopeful element to bring some semblance of order back in the utterly disrupted fabric of our society.

In a growing number of books on the topic of algorithms and inherent flaws, this lucid and insightful volume rises above the others.

The authors Michael Kearns and Aaron Roth, both computer scientists from the Computer and Information Science department at the University of Pennsylvania, contend that allegedly anonymized datasets routinely leak our private information and that supposedly fair statistical models end up with significant racial and gender bias.

Furthermore, legal and regulatory bodies are woefully behind in a relatively flat trajectory when the algorithms and artificial intelligence have been exponential in their development and deployment.

As a possible solution to this imbroglio of injustice and technology, the authors suggest a set of principled solutions based on the innovative socially-focused algorithm design that essentially embed human (and humane) principles into machine code while concomitantly support this exponential rise in machine intelligence.

The authors are able to write the content of this volume right in the small but delectable sweet spot between technological substance with the appropriate jargon and the social implications of all of this paradigm.

In addition to chapters on algorithm privacy, fairness, and manipulation, there are several very memorable sections accompanied by intriguing titles such as Jump Balls and Bombs and Power Poses, Priming, and Pinot Noir.

Perhaps our authors as well as Alan Turing are prescient and correct in their general philosophy that “machines, not humans, should deal with machines”.