ARTIFICIALLY YOURS

Can machines imitate friendship?

Ethicist and educator Tiberius dives headfirst into the philosophical puzzle of chatbot companionship and what it reveals about the nature of friendship. Concerned with ethics, moral psychology, and well-being, she approaches the subject not as a technologist but with a philosophical interest in what makes friendships tick. Initially skeptical of “chatbot friends,” she acknowledges that human friendships themselves vary widely in quality and depth. The book’s central aim is less about predicting technological change than about understanding what human friendship is and how it affects our well-being. Early chapters define the conceptual groundwork. Machines can have goals—much like a chess program that aims to win—and intelligence can be understood as the ability to accomplish complex goals. But consciousness is the ability to have a subjective experience, which contemporary chatbots lack. Tiberius challenges narrow definitions of friendship—real friendships exist in many forms and degrees, and even imperfect friendships have value. Chatbots can mimic some features of friendship—empathy, attention, companionship—which is part of their powerful appeal. Although talking with chatbots can help people cope with life’s challenges, they cannot genuinely admire us, share our vulnerability, or understand human suffering and mortality. Can future superintelligent chatbots self-improve and overcome their limitations? Tiberius observes, “I don’t see any reason to trust that AI left to learn on its own will learn what we want it to.” The profit motives of AI companies also give her pause. Ultimately rejecting both techno-pessimism and techno-optimism, she argues for a middle ground: AI companions may offer comfort or teach us important skills, potentially enriching and supplementing the irreplaceable value of human friendship.

An accessible, clear-eyed look at AI companions.