Caitlin Green

To me, that's a really good group because you're in a new city potentially, meeting new people.

You're potentially in a dorm.

It's a time when people are really lonely in life.

Is there any other good reasons we might want to look at first-year university students when testing out loneliness stuff like this?

The human peer, I imagine.

But there's kind of an issue here.

Like your results, they're not promising for chatbots.

Like if you're showing that they don't actually reduce loneliness over time, people are still going to use them when they are feeling lonely.

Is there a way to design them better?

If we just made some changes, maybe they actually would help.

Maybe some prompts saying, hey, I know you're talking about this bot about feeling lonely and that's okay, but you should also talk to a human.

Like any recommendations for how these things should be designed based on your finding?

I wonder if you programmed a bot to like accept flowers or accept a gift, if that would help.

It's kind of a ridiculous example.

Where would you even send the flowers?

ChatGPT's headquarters?

I don't know.

Well, Ronnie, I've enjoyed the chat.

Thanks for sharing your research with me.

Thank you.