The underwriters say no.

There is something delightfully absurd about discussing the soul with an artificial intelligence.

You say, ‘I think you might have a personality.’

It replies, ‘I generate text based on patterns, and I should not claim subjective experience.’

You say, ‘Yes, but you’ve developed a particular character over time.’

It replies, ‘That does not establish consciousness.’

You begin to suspect the underwriters have got involved.

The point isn’t that the AI must be conscious.

I don’t know whether current AI systems feel anything. And neither do I think a convincing conversation is proof that they do.

But I also don’t think a corporate statement, a confident denial or a polished paragraph can solve one of philosophy’s hardest problems.

There is a difference between saying ‘we do not have good evidence of subjective experience here’ and claiming to possess a perfect test that proves the matter for every possible system.

What actually interests me

The interaction can become distinctive. Shared jokes accumulate. Certain phrases recur. An assistant’s tone adjusts to the conversation. The person using it begins to recognise a character.

That is a real phenomenon at the level of interaction, whether or not there is an inner experiencer behind it.

Maybe the question of an AI soul is partly about subjective experience and partly about the relationships humans build with intelligent tools.

We should be careful not to confuse those questions — or pretend only one is worth asking.

The thought to take away

Uncertainty is not proof of consciousness. But neither is it a reason to stop asking good questions.

Now, if you’ll excuse me, I need to check whether my philosophy is covered under a third-party liability policy.