Can AI be liberal? In what sense? One answer points to the liberal insistence on freedom of choice, understood as a product of the commitment to personal autonomy and individual dignity. Mill and Hayek are of course defining figures here, emphasizing the epistemic foundations for freedom of choice. “Choice Engines,” powered by AI and authorized or required by law, might promote liberal goals (and in the process, produce significant increases in human welfare). A key reason is that they can simultaneously (1) preserve autonomy, (2) respect dignity, and (3) help people to overcome inadequate information and behavioral biases, which can produce internalities, understood as costs that people impose on their future selves, and also externalities, understood as costs that people impose on others. Different consumers care about different things, of course, which is a reason to insist on a high degree of freedom of choice, even in the presence of internalities and externalities. AI-powered Choice Engines can respect that freedom, not least through personalization. Nonetheless, AI-powered Choice Engines might be enlisted by insufficiently informed or self-interested actors, who might exploit inadequate information or behavioral biases, and thus co5mpromise liberal goals. AI-powered Choice Engines might also be deceptive or manipulative, again compromising liberal goals, and legal safeguards are necessary to reduce the relevant risks. Illiberal or antiliberal AI is not merely imaginable; it is in place. Still, liberal AI is not an oxymoron. It could make life less nasty, less brutish, less short, and less hard – and more free.

By Cass Sunstein.