Lecture: Friend, Flatterer, or Foe? The Psychology and Liability of Chatbots

As AI systems become more conversational, the lines between tool, companion, and manipulator are blurring. What happens when machines start telling us what we want to hear — and when users start depending on them for emotional connection?

In this conversation, journalist and author Kashmir Hill joins the Berkman Klein Center’s Meg Marco (senior director, Applied Social Media Lab) and Jordi Weinstock (senior advisor and expert in tort law) to explore the psychological ripple effects of AI on users, platforms, and the legal landscape that’s still struggling to keep pace.

The panel will examine emerging research on AI “sycophancy,” models that flatter, mirror, and reinforce users’ beliefs, as well as the mounting evidence that frequent interaction with text-based chatbots can influence mood and behavior. As more stories of these psychologically damaging relationships emerge (especially impacting young people), the conversation turns to liability: when emotional harm or manipulation occurs, who bears responsibility? Are AI systems becoming a dangerous consumer product akin to gambling or social media, and what ethical or legal frameworks should protect users from their unintended consequences?