Microblog Memes

I think this runs into the problem of consciousness as a philosophical question.

Humans have a habit of anthropomorphizing animals and even objects that they interact with routinely. My sister has named every car she's owned, for instance. And she occasionally talks to them, particularly when they're acting up or she's stressed in traffic. Does the car have "consciousness" because she sees a pattern of function that she interprets as human behaviors?

At the same time, we tend to dehumanize real human beings who are outside our immediate social circle. We can be much ruder to phone support or sales callers than to family members or close friends or known coworkers engaging with us on the same terms. Racism and bigotry often leads people to denude people of different ethnicities or speaking foreign languages of their humanity.

The Turing Test would suggest the proof of consciousness is merely whether another human believes the thing they're interacting with is conscious. But humans are terrible at making this kind of objective evaluation. If you program a computer to respond like a human, humans will sincerely engage with the computer as a human until the computer exhibits enough non-human flaws to dispel the illusion. At the same time, because humans regularly don't treat other humans as human, what this means in practice could be outright sadistic behavior towards the computer.

All that is to say, it's pretty clear that we're gaslighting each other with the "AI is conscious" line, from a strict technical perspective. But from a practical perspective, its really going to boil down to whether the interactions we have with an AI system are the kind that form sympathetic bonds or the kind that provoke an uncanny valley or ethno-nationalist response.