Sycophancy or Small Talk?
Can AI Be A Friend?
I don’t usually watch South Park, but this week I did. The episode was called “Sickofancy.” In it, a character treats “ChatGPT” almost like a confidant. His wife gets so annoyed she starts mimicking the chatbot. It’s hilarious, sharp satire, but it also unsettled me. It captured, in a cartoon way, both the appeal and the emptiness of AI as a “friend.”
That struck a chord because the same question keeps popping up in my own life: is AI your friend? Can it be? Should we even talk about it in those terms?
Why this is on my mind
I research anthropomorphism and trust in large language models, in systems like ChatGPT, Gemini, Claude. I hear this “AI as friend” question in academic spaces too: I was just at a panel that asked it directly. But what stands out to me is how often it comes up in casual, everyday settings. People want to talk about it at parties, on walks, over coffee.
A commute story that stuck with me
Someone told me their son no longer listens to music on his long commute. Instead, he chats with ChatGPT. Just small talk, just passing the time. A companion on the road. I haven’t done this myself yet, engage in small talk with a chatbot I mean, but I plan to try it and see what it feels like.
Questions I’m puzzling over
What do we actually mean by “friend”? Is it emotional reciprocity, moral understanding, reliability…or something looser, like “keeps me company”?
If people feel friendship with an AI, does the lack of real reciprocity (the “give-and-take” of human friendship) matter?
Is anthropomorphism inevitable? We slip into human terms to make sense of non-human systems. When does that help and when does it become harmful?
What are the practical risks and benefits? A commute chat feels light and harmless; other cases might be manipulative or misleading.
Is corporate ownership a dealbreaker for AI friendship?
Invitation
I don’t have conclusions here, only questions and curiosities. I’m going to try the “AI commute chat” experiment this week and report back. Have you ever used AI this way? Do you call it a “friend,” even jokingly? If so, what does that mean to you? I’d love to hear.
Once I noticed it, I saw it everywhere. Not just in what AI says, but how it says it: fluent, friendly, polite, confident. These systems don’t just generate language; they perform credibility. And we’re wired to reward that.