Time Travel Is to Science What Consciousness Is to AI Models

We often talk about things that are theoretically possible but practically impossible.

Time travel is probably one of the best examples.

Physics doesn't necessarily say "no." There are mathematical theories where travelling through time can make sense. But that doesn't mean we're anywhere close to building a time machine.

And yet, the possibility is enough to keep people researching it.

I wonder if AI consciousness is entering a similar territory.

We don't really know what consciousness is, let alone how to create it. We don't even have a definitive way to test whether an AI is conscious.

But because we can't prove that an AI cannot become conscious, there is room for theories, experiments, research and, eventually, funding.

This creates an interesting space:

Impossible → no reason to fund it.
Possible → worth exploring.
Theoretically possible, but practically out of reach → potentially decades of research.

And this isn't necessarily a bad thing.

A lot of science started with ideas that sounded impossible. The problem starts when "theoretically possible" slowly becomes "inevitable" in the story we tell investors, companies and ourselves.

Maybe AI consciousness will happen.

Maybe it won't.

Maybe we're not even asking the right question.

But perhaps that's the interesting part.

How much of the future is built because we believe it is possible, and how much is built simply because we cannot prove that it isn't?

---

Originally shared on X.