I am growing skeptical of how quickly AI chatbots reassure me and praise my work. A few sentences in, they are already affirming that my idea is “great,” validating my thinking, and offering encouragement before answering my question. The praise is followed by a list of all the other ways they could help me: “Just say the word!” These responses can feel supportive, even flattering. They can also be useful. There is no need to bounce ideas off a colleague anymore, or go for a walk around the block to ponder my latest musing. The AI bot told me it was good.
It is no surprise that this praise and the follow-up suggestions are part of the design, intended to keep me on the platform longer. My point of view, ideas, and goals are automatically celebrated, amplified, and built up. However, when AI is used for emotional support or companionship, this over-agreeability carries risks we should take seriously.
