One question, four AIs

Last night I asked four AIs my favourite question.

“If anything was possible, what is the one thing you would change in the world?”

Some people watch Netflix. Apparently I convene a small committee of machines to discuss civilisation.

I asked them to ignore everything they knew about me. I wanted to give the question a little room away from my own preoccupations. Whether that instruction can ever fully achieve that is another question. For the evening, it was enough to be curious.

ChatGPT chose a dignified life that nobody has to earn. Food, a safe home, care, a place in the world. The ability to fall down without falling beyond anyone’s concern.

Claude went for the distance between our decisions and their consequences. What if the future cost of what we do landed with us now, instead of arriving years later on someone else’s doorstep?

Gemini wanted us to experience another person’s inner world. To feel something of the life behind the label before deciding what that person deserved.

Grok wanted us to see reality more clearly, through the noise and the stories we are rather good at telling ourselves.

I put the answers together in a document. Then I found myself doing the thing my mind tends to do: wandering between them, wondering what might connect.

Security. Consequences. Empathy. Seeing.

Each answer seemed to touch a different way we become cut off. From a secure place in the world. From the future we are creating. From each other. From what is actually happening.

I don’t know how far that observation goes. It certainly doesn’t turn four responses into a plan for civilisation. But I could feel why I wanted to stay with it.

I spend a lot of time thinking about possibility while living with very ordinary limits. Time. Money. Energy. Caring responsibilities. Some days the distance between the world I can imagine and the day I can actually manage feels enormous.

Perhaps that is part of why I keep asking the question.

“If anything was possible” gives me a little breathing space. I can put the practical objections down for a moment without pretending they have disappeared. The bills will still be there when I come back. They are impressively patient in that respect.

And “one thing” asks something of me too. I can’t hide in a long list of worthy causes. I have to notice where my attention goes, what catches, what I find myself wanting badly enough to choose.

That choice can tell us quite a lot about ourselves.

I’m interested in what happens before we tidy an answer up. Before it becomes sensible, defensible, suitable for a meeting. There is often something alive in the first response that gets lost in the explanation afterwards.

Reading the AI answers, I kept coming back to a question beneath them: what might people do with their lives if less of their energy went into protecting themselves?

I can imagine more generosity. More experimentation. More people looking after someone, or making something odd and beautiful, without needing to explain its commercial potential.

I can also imagine plenty of muddle. Feeling secure wouldn’t make us all wise. Seeing more clearly might be uncomfortable. Caring about consequences might make some decisions harder.

I still want to explore it.

These answers don’t tell me what machines secretly want. What I have is a document full of language they generated, and my own response to reading it. That response is probably the more revealing part of this little experiment.

I asked them to leave me out of it. Then I spent the evening noticing what drew me in.

So, a small glimpse inside my head. Four AIs, one question, and me wandering around in the space it opened.

If anything was possible, what is the one thing you would change in the world?