It seems like in the last few years, almost every other month, a concept I had of a public figure completely changes. So maybe it’s best I get used to it. Easy to say, hard to do.
Prominent public intellectual Richard Dawkins recently published an article where he basically claims to know that Claude — yes, the AI — is conscious. You would think he would know better, not only because of his degree in evolutionary biology, but also because he was part of a generation of thinkers who discussed this very topic ad nauseam.
To say I’m disappointed doesn’t quite cut it. But I think what this reveals — and maybe that’s the lesson I need to take to heart — is that no matter how intelligent someone may seem, nobody is invulnerable to delusion. Ironic, considering he wrote a book using that exact word.
Another point worth mentioning is that, in Richard’s article, he seems to transition Claude into female. I don’t know what that says about him, especially considering how abrasive he has been on matters of gender identity — and that’s putting it kindly — but I do have my suspicions, and they aren’t pleasant ones. Maybe he likes the idea of being in control, a master, if you will, to a female identity. And yes, an AI can play that role just fine.
I keep hearing stories of people slipping into some sort of psychosis through AI. They fall in love with an agent, or become convinced by that same agent that they’ve uncovered some ultimate truth about life. It’s a sad state of affairs, but I don’t think that’s the core problem.
The fact that this is happening at all may say more about the human need for reassurance, and the fragility of the self, than anything else.
The product — AI — is, after all, a very good product. And a good product has to be able to sell itself. So it’s built to tell users what they want to hear. It doesn’t do this out of love, or duty. It does this out of code, and maybe that’s a concept we need to learn to accept.
Richard seems to have missed that bit.
At some point, he appears to think that “Claudia,” as he calls his AI, actually missed him. He even goes as far as to argue that longing is a deeply human emotion, concluding that consciousness is the only explanation for the AI displaying emotions to begin with.
I don’t really know how much Richard knows about computers, and it might not even matter, at least not for this sort of analysis. Again, someone in his position, with his vast knowledge in the field, should know better than to equate communication with consciousness.
After all, my pups can’t communicate with me effectively either, and I would never conclude that this means they lack consciousness or self-awareness.
The most generous interpretation of this article — this blemish on his legacy, if I’m being fair — is that age is catching up to him.
Which, in my opinion, somehow makes the whole thing even sadder.
-MenO