As a side hustle (though more as a way to explore AI from yet another angle), I have begun rating AI for an online company. For something like $20-$30/hour, they present me with a variety of responses to different prompts, and I then evaluate those responses for what was more effective, ranging in topics from the way it handled complex emotions to simply “did it answer the prompt?”
Doing this for a little bit has helped me see more clearly what the limitations of AI are right now. It really struggles in some areas (like . . . simple arithmetic) and excels in others.
But beyond the observations I’ve been making about AI, I’m also seeing the prompts users are turning to AI to answer, and it’s giving me a fair bit of insight on humanity, as well. Most of those insights are sad and depressing.
Yes, people turn to AI for frivolous things or silly things or obvious things, but the other week, I was rating voice interactions where people had given AI some sort of a vocal request, and I was to develop some criteria for how AI ought to respond. What sort of tone should it have? What sort of pacing? How about word choice? (For a techie ex-linguist, this is a veritable goldmine of things to think about.)
Time and time again, the prompts were bleak. People were asking AI how to make friends. How to lose weight. How to handle difficult discussions. Just a veritable sea of loneliness and despair.
The thing is, I remember my time as a missionary talking with people on the street day in and day out, doing my best to help people. Yes, I get that there are some real alarm bells that go off for people as soon as someone starts talking about religion, but even taking that into account, I saw people consistently portray themselves as “all set.” They didn’t need any help, thanks very much. They were perfectly happy.
Never mind the fact that East Germany then was really struggling. That unemployment was a real concern. That the future of the country was all over the place. People just don’t like others to see them as anything but self-sufficient.
But as soon as they can talk to something that doesn’t judge them at all? That is guaranteed to respond as helpfully as possible? Some people seem to be much more comfortable revealing how they really feel. Yes, I get that many people might not feel that bad. But . . . I don’t really believe that, to tell the truth. I think there’s an awful lot of angst, confusion, and loneliness out there.
And they’re turning to AI to try to solve it.
Look, I’m a huge proponent of AI. I turn to it for all sorts of reasons. But I’m also a real believer in the power of personal connections. There are some things AI just can’t replace. It reminds me in some way of the Sondheim song, Being Alive:
Yes, people can be very challenging at times. Even people we love and who love us can stretch us and test us and make us lose our patience. But those same interactions are also the ones that make life worth it. They help us become better people.
I suppose the good news (for now) is that people were turning to AI for advice in how to handle other people. In other words, they hadn’t completely abandoned the idea of interacting with actual people yet. Then again, I don’t know how long that will stay the case for some people. AI is very good at telling us what we want to hear. At reassuring us that we’re smart or funny or important. AI doesn’t like to challenge you on anything, and I imagine that could prove to be very appealing.
Do I think this is something that’s going to shut down social interaction across the world? No. But I do think it could significantly influence it in some circles, and that’s something to be very cognizant of. The more we splinter, the more difficult coming together becomes.