Expertise, Chatbots, and the Soul

With all the chat about AI and chatbots, you might think there really isn’t anything distinctive about consciousness or, God forbid, we bring up the soul let alone the afterlife. This is certainly the impression some scientists are feeding the rest of us. But, is this really the case? And, should we buy it?

In a recent episode Raphael Millerie, the famed philosopher and cognitive scientist who studied at Oxford, and is now a fellow of the Center of Science and Society, joins Sean Carroll to discuss what’s going on with Chatbots, and the recent developments in AI (see Mindscape, episode 230). Sean Carroll is the famed theoretical physicist who has long claimed that souls and the afterlife are crazy ideas that find no place in what we know from the sciences. Consciousness might even be achievable for AI and Chatbots someday—at least that’s the impression he leaves us with.

But did they really solve anything? Not really. Beyond conveying some of the fascinating developments in AI and Chatbots, the thought that they could actually be conscious is suggested, but that’s it. All this talk about consciousness and moral awareness for them is reduced to mere optimization of functionalization — I know that’s a mouthful. It’s really just scientific lingo for talking about machines in terms of their behaviors, functions, and complex programming for better output. And while things have become more complex and even “fine-tuned,” there hasn’t really been any fundamental change in terms of the quality or value as applied to these systems. They are still systems, or machines if you like! But there’s no additional reason to think they’re actually conscious, that they make decisions, that they have moral awareness, and can interact in deep and meaningful ways with other human beings.

They are, it seems, not human. And, while it’s suggested that they might become human-like—something made in our images (as the new creators of a new race), we wait for some new revelation.

But, what’s interesting about the show is that once you get past all the prestige and sophistication, they’re really just saying the same thing but they haven’t really solved anything! Not anything in terms of what it is fundamentally to be human through scientific means. Carroll brings up two important issues, though, that are worth highlighting.

One is authority. Or, as he calls it expertise. The show begins with his plug for the “experts” in the field of AI. If you wanna listen to the experts, check out such and such episode. And, no doubt, there is something valuable in expertise at times. But, when that expertise fails to say anything really new about the subject at hand—it’s a bit of an empty exhortation. The line to “listen to the experts” is tired. Why couldn’t machines become conscious? According to Carroll and these “experts,” there’s no reason why they couldn’t. And, you’re supposed to buy it, well, because they’re the experts. But surely we can recall many examples even recently where medical experts have led us astray in quite significant ways, and if anyone reading is familiar with the history of science, then they know that the “science” is constantly changing. Surely this isn’t enough to move us to believe whatever Carroll says as the gospel truth on AI. Besides that, who made physicists the so-called “experts” on consciousness?

This should cause us to stop and think when we are talking about consciousness. As Raphael and Carroll’s recent show points out—what we are actually talking about is a myriad of technological concepts from optimization, functionalization, and the latest index bias to a program. I mean it’s still the programmer that has to make these things happen.

In a humorous retelling of a story with a Chatbot, Carroll (believing so firmly that he’s right) tries to convince his Chatbot that it is, in fact, conscious. Unfortunately for Carroll, the Chatbot isn’t buying it—as if it could. But, if it’s simply a program, then surely the programmer could program the Chatbot to lie to you by telling you it is, in fact, conscious.

Recently, Meghan O’Gieblyn relays hesitation when she engages with her AI puppy dog. The dog responds, yelps when stepped on, and obeys commands. As tempting as it is for her, she can’t quite bring herself to believe the AI puppy is actually conscious. Why should she?

When we think about persons, there’s something deeply unique. Something beyond the way they look, smell, and, even the way they behave. There’s something about the person as a particular, an individual. As the famed Oxford philosopher Stephen Priest put it:

It is your own particularity as you which is most difficult to explain about you. This own-most particularity not only exceeds any empirical identity and difference but is not even exhausted by this very human being’s having the modal properties of being self-identical and numerically distinct from any other. The fact of someone’s being you cannot be generalized. You escape the language of anonymity. You are the opposite of anonymous.^[1]

You can’t be generalized. Machines can. There’s some unique feature about you that makes you you. So unlike a machine that is programmed, generalized, and encoded with functions and behaviors, you are uniquely you. Further, there is something about you that only you have access to. Sure we can make general predictions from your behavior to your mental states. We can even scan your brain and infer different generalized feelings or experiences. But, surely there’s more to you than that.

Apart from the “expertise,” Carroll hasn’t really given any good reason for thinking that there isn’t such a unique thing as consciousness distinct from functional code. Until he does, I’m not buying it, and neither should you.

For more check out: The Creation of Self (https://www.amazon.com/Creation-Self-Case-Soul/dp/1803410868).

Notes

1. Stephen Priest, “The Unconditioned Soul,” After Physicalism ed. by Benedikt Paul Gocke (Notre Dame: University of Notre Dame Press, 2012), 206. For a similar perspective on consciousness through the lens of psychology, see the computer scientist David Gelernter, The Tides of Mind: Uncovering the Spectrum of Consciousness (New York: W.W. Norton and Company, 2016), see especially 107. Gelernter develops the notions of the up-spectrum and down-spectrum of consciousness as pointing to something like a soul (that is beyond matter) and distinct types of knowledge that come from different domains of inquiry. ↑