Q&A with Dr. Cass Sunstein on new book “Algorithmic Harm”

Drs. Oren Bar-Gill and Cass Sunstein have a new book out called “Algorithmic Harm: Protecting People in the Age of Artificial Intelligence.” It details the harms and benefits of AI and algorithms and discusses the different ways in which algorithms are being or may be used in consumer and other markets. CHIBE spoke with Dr. Cass Sunstein about the book.

Your book describes the many ways in which algorithms in consumer markets (and other markets) are causing major harm. How do you cope with how depressing this situation is, or do you feel optimistic that policy/regulatory responses can help address some of the issues?

In my view, optimism and pessimism are less constructive than analysis. One of the goals of the book is to analyze the sources of algorithmic harm and then to see what can be done about it. Providing people with relevant information is one reform, so that they know what they are doing. Algorithmic transparency is another, so that people can know what algorithms are doing.

Can you give us an example of how a company’s AI could take advantage of people who suffer from optimism bias or present bias?

Sure! Suppose that people are optimistically biased. You could market a hair restoration product to people who think it will work, when it really is unlikely to work. I myself might buy that product, because I am optimistically biased (and could benefit from it).

What’s one takeaway you’d want people to come away with after reading your book?

The most important takeaway is that any book that has one takeaway is a pretty terrible book.

The second most important claim (not a takeaway) is that algorithms, including personalized algorithms, can be great for consumers who have adequate information and do not suffer from behavioral biases – but that when personalized algorithms can figure out which people lack information and suffer from a behavioral bias (present bias, availability bias, optimistic bias), there can be a world of hurt out there.