Online retailers like Amazon may soon be able to predict what consumers want before they know themselves, said Harvard Law School professor Cass Sunstein, who served on the National Security Agency oversight panel. With shoppers’ personal data, retailers can create interfaces that reflect individual preferences, even potentially shipping items before they are purchased.
In a lecture on Thursday afternoon, Sunstein explored the advantages and disadvantages of impersonal default laws, generic data rules applied en masse, versus active choosing, where the consumer decides which information to share with retailers. He described a hybrid called personalized default rules, which allow the entities that consumers interact with to know their preferences and adjust their interface accordingly.
Sunstein served as administrator of the White House Office of Information and Regulatory Affairs from 2009 to 2012. He said his primary focus in this position was making sure that the benefits of federal regulations justified their costs.
“The president charged me with ensuring that our regulatory system was compatible with his goal of getting out of a very tough economic hole,” he said.
This government experience sparked interest in default settings, active choice and the growing concept of personalized defaults. He said people must be able to control their own privacy settings, but the task of constantly managing these settings would be burdensome and could take away from more important obligations.
“It’s important to have active choice, where people decide individually what they feel to be best for themselves,” he said. “However, if people had to make active choices for every aspect of their lives, such as minute phone and computer settings, we would quickly be overloaded.”
As data mining becomes more prevalent, having personalized defaults, or defaults suited to individuals in particular, is increasingly possible, he said. These personalized settings, however, face three main criticisms.
Gathering accurate personal information can be difficult, may raise qualms about personal privacy and creates the risk of people not developing their preferences over time. People may not evaluate their own preferences because they would be limited by their initial default settings, potentially decreasing shoppers’ adventurousness in purchases such as books and music.
“You want an architecture of serendipity, not an architecture of control,” he said.
He said that despite these qualifications, personalized defaults are “the wave of the future.”
Clara Wang ’17 said she enjoyed the lecture, but she expected Sunstein to speak to his NSA background.
“I was a little disappointed because part of the advertising for his lecture suggested it would be about the NSA, so I thought he might address Edward Snowden’s whistleblowing or NSA surveillance,” she said. “Nevertheless, I thought it was pretty interesting and he had some different ideas on how ‘opt-out’ policies might be more effective.”
