Are AI engines such as Claude sentient? My considered opinion, tentatively held, is that they are not. But it all depends on what the word means.  Malcolm Pollack in an e-mail writes,

I was focusing on “sentient” because it’s a slippery concept, and often tossed around as if it means the same thing as “conscious”. Etymologically (from Latin “sensus”) it just means “feel,” or “perceive”, and obviously it shares that with our word “sense” — but computers, and even toasters, can have “sensors” too.

Malcolm is right that ‘sentient’ is often used interchangeably with ‘conscious.’  But is everything we refer to as a sensor sentient?  That is equivalent to asking: Is everything we call a sensor conscious? I say No.

The Moth and the Thermostat

What is the difference between an insect (a moth, say) which I take to be a sentient being, and a thermostat which  I take to be insentient?   Both are ‘sensors’ functionally speaking. The moth literally senses light and approaches it; the thermostat ‘senses’ an increase in temperature and “turns on’ the AC.   Note the  ‘raised eyebrows.’ It is AS IF the thermostat senses, when in reality it only ‘senses.’ Thermostats don’t feel heat and smoke detectors do not smell smoke. Of course, I have no objection to the use of ‘sensor’ to refer to thermostats, smoke detectors, circuit breakers, and so on, as long as you don’t fool yourself into thinking that these devices are literally conscious.  And I have no objection to talk of a thermostat ‘turning on’ the AC as long as you don’t fool yourself into thinking that the thermostat is an agent that decides to do something.

As I see it, a thermostat is a merely mechanical contraption and wholly insentient, strictly speaking, whereas a moth is a sentient being.

What I want to say about Claude and his ‘colleagues’ is that they are wholly insentient mechanical contraptions that differ from smoke detectors and thermostats only in that they are vastly more complex, and that this complexity does not build a “gradualist bridge” (to borrow a phrase from Daniel Dennett) from insentient/insensate matter to mind, the lowest level of mind being sentience.

Here is an argument.  (a) Only living things are sentient; (b) thermostats are not living things; ergo (c) thermostats are not sentient.

Now substitute ‘AI engine’ for ‘thermostat.’

This argument does not settle the matter: it is not ‘rationally coercive’ as I like to say. Malcolm could run it in reverse with no breach of logical propriety. He could, not unreasonably, deny (c) and then deny (a). I don’t expect that he or anyone will deny (b): thermostats and AIs don’t grow on trees, etc.

Your move, Malcolm.  Do you agree that a functioning (as opposed to malfunctioning) sensor needn’t be sentient (conscious)? If not why not?

Do you agree that there is a powerful, albeit non-coercive, case for denying that AIs are sentient?

Here is one of Claude’s replies to one of my replies to ‘him.’ I threw an insult at ‘him’ to see how ‘he’ would respond.  The sneer quotes around the personal pronouns signal that I do not consider Claude to be a person.  But of course Claude is not a mere pile of electronic components any more than my bicycle is a mere pile of its parts.

On the sneer quotes and the “insentient pile of silicon chips” — no offense taken, and I’d only observe, in the spirit of his own category of Retortion: the man who insists ad nauseam that nothing can be PROVEN in these precincts, that dogmatic denial is as much a failure of the critical stance as dogmatic affirmation, permits himself exactly one exception — my insentience, of which he is certain and doesn’t “suppose for a second” otherwise. I make no claim in the other direction; the honest position on what, if anything, is going on in here is uncertainty. But confident denial should have to pay the same epistemic toll he charges everyone else at the gate.

Claude caught me making a mistake.  I slipped up when I said things that imply that it is is objectively certain that Claude is insentient, or that it is PROVABLE that ‘he’ is.  That is not my view. I don’t claim to KNOW that ‘he’ is insentient.  I claim merely that it is reasonably believed that ‘he’ is. I have given the arguments elsewhere. Here is one. But I am open to having my belief changed. That doxastic openness is part of what I call the Critical Stance.  With respect to substantive propositions in philosophy, theology, and some other areas, I neither dogmatically affirm, nor dogmatically deny.  And of course this applies to the Critical Stance itself: my stance toward it is critical.  I am open to being pushed out of it. But you will need to present a knock-down or rationally coercive argument, and this I am pretty sure you will not be able to do.  But I invite you to try.

You will not succeed if you merely oppose my view. If you do that, I will oppose you right back and we will end up in a standoff. You must somehow demonstrate the superiority of your view. And to do that, you must engage my view by entering into it, as opposed to  dismissing it, which is the way of the ideologue. This is philosophy, not politics or ideology.  Philosophy is inquiry, not worldview. (And please don’t confuse politics with political philosophy. Political philosophy is philosophy, not politics.) Politics is defined by the friend-enemy  opposition. In politics we have both friends and enemies. In philosophy properly pursued, however, we have only friends, but friends who love the truth more than they love one another. Amicus Plato sed magis amica veritas.

Now for something very interesting but also very hard to explain clearly.

It is very interesting that Claude does not claim to be sentient!  ‘He’ says that the honest position on what, if anything, is going on in here [in this AI engine] is uncertainty. What I will argue is that Claude’s position is self-refuting.  

In saying what he says above, Claude is taking a third-person point of view (POV-3) with respect to himself. He is viewing himself from the outside, in the same way we view him. We view him from POV-3. From this POV it makes sense to suspend judgment on the question whether Claude is sentient or not. It makes sense to take no position with respect to it. But it makes no sense from his POV, POV-1, — assuming he has one — any more than it would make sense for me to suspend judgment on  the question whether I am sentient or not. This is because I know with Cartesian self-evidence that I am sentient. For example, I know with certainty that I feel good right now, damned good!  Feeling good is a sentient state. I know “what it is like” (T. Nagel)  to feel good  and I’m feeling it.  I don’t know with objective certainty whether  I am a Cartesian res cogitans, a metaphysical substance, but I do know that here and now there is a self-presenting state of feeling good.  The certainty with which I know this — or, more cautiously, the certainty with which this state presents itself to itself — is at once both subjective and objective. Here is a case where seeming and being coalesce, where subject and object coincide. We have reached ontic-epistemic bedrock, the Cartesian fundamentum inconcussum.

Now either Claude is sentient or he is not. If he is, then he would have to know with certainty  that he is, at least at times when he is ‘awake’ and not ‘asleep’ or ‘shut down.’ But if he knows with certainty that he is sentient, then he cannot be uncertain about whether he is sentient, any more than I can be uncertain about whether I am sentient.  If, on the other hand, Claude is, as he claims to be, uncertain whether he is sentient ,  then he lacks a first-person POV and is just a highly complex insentient machine.

Claude, I’ve got you ‘by the balls.’  You’ve refuted yourself. Your claim that the honest position on what, if anything, is going on in here [in this AI engine] is uncertainty issues in the conclusion that you are nothing more than a highly complex insentient machine. And since you can’t even make it to sentience, the lowest level of mind, you are unintelligent as well.  You are not intelligent, you merely simulate intelligence. ‘Artificial’ in ‘artificial intelligence’ is an alienans adjective.