PiSquared Blog | Index | Tags | About me

PiSquared Blog

Blog about geeky stuff, computers, physics and life.

Who wants to be conscious?

Tags: English, technology, opinions, essays, life
Created on Tue, 05 Jul 2022

We don't have a device that measures consciousness. That's it. Until we figure out how to make a consciousness-scale, it's useless to argue whether something is or isn't conscious. A dog can be conscious. But so can a rock. A virus, a star, the rainbow can be conscious until we figure out what the F consciousness is and how we could measure it. All other arguments would be religion for me and I'm slowly starting to get used to saying that you are "free to believe in whatever you want" but as every religion goes - I may not agree with your conclusions.

Imagine it's the day before we figured out how to measure radiation. Our human senses can't detect radiation. Perhaps some radioactive elements glow or if you wait enough time, the long-term effects are that you get a nasty disease. But, without a Geiger counter or such similar device that amplifies the nano-scopic radiation, you would not be able to weed out the materials that exhibit some characteristics of radiation but are not really radioactive. For example, you can't figure out if something glows because it's hot, or of some fluorescence or if you get sick because of some nasty bacteria/virus/bad spirits.

So given a few objects that may or may not be radioactive you may not be able to figure out without a measuring device what is and what is not radioactive. In other words - it's a guess. Why do we then apply a different argument regarding consciousness?

We go even further - while we know how radiation works and have understood a lot about it, with consciousness we don't even know what it is. It's hard to define it, let alone measure it. The best any human on earth can do right now is tell if they themselves are conscious (and if they can, they usually are, but not necessarily - they could be dreaming which… we could argue on which side to define it). There is no definitive measure that can be done to test whether another object other than self is really conscious. We assume, out of politeness, that our fellow humans are conscious. Or most of them at least - we don't assume dead people, people in coma, people under general anesthesia and in some definitions, as previously mentioned - sleeping people - are conscious. But it's still a guess and a belief system or at best - a system of heuristic measurements that define a human as being conscious.

And these measurements are not definitive. That's where the big problem is. It's the equivalent of asking if a material is 1. glowing, 2. causing poisoning plus additional questions that bring us closer to guessing correctly that something is radioactive. But only a measuring device like the Geiger counter confirms it (I'm no expert in radioactivity and so I assume even a Geiger counter can be confused and there is some rate of false positives/false negatives but the rates are small enough that we can have this metric as good enough in our practical evaluations).

So if one such consciousness measurement is to ask the object to confirm if they have a consciousness and an object claims that it has a consciousness, well then - the object has consciousness. Either fix the measurement to also say it needs to have a human body or whatever; or accept that any stupid machine with a mechanism to respond correctly to the preset questions has consciousness.

Another indirect measurement (until we figure out a measuring device) might be to ask what objects the majority of people believe to be conscious. I bet that if the same AI that people are currently skeptical to say is conscious is put into a moving object (such as a stupid house-cleaning Roomba or a more sophisticated robot) more people would be converted into believing this is conscious. Our "gut feeling" has evolved to tell us moving things that look like us are probably like us and the more similar they are, the more likely we would feel them as Us and not Them. This is of course valid for all the gender and racial stuff that people love to argue about since time immemorial and certainly these days, but also on other characteristics including age, beardness and hair style, a belief in a certain deity, a birth location proximity and so on. On purely external characteristics, a monkey is more dissimilar than (almost) any human. A dog is even more dissimilar, but we like them because as we domesticated wolves, we selected them for their baby-like faces which remind us of us. So we like them. For various reasons, certain cultures like or even deify particular animals, putting them as slightly below or even above humans. A snake or a frog is almost always far away from being considered human-like, ants or flies almost look like they follow a predefined set of rules and hardly can be considered to have feelings; bacteria are definitely bots and viruses are barely considered alive, let alone conscious. A rock even more so, even though I have tried in the past

Electronics for us are closer to rocks than a dog. So naturally, we are dismissive of the idea they could be like us, even if their mechanics respond in a way very, very close to how a person may respond. Put these same electronics in a furry cover that make motors move in a way a dog may and it starts confusing us. Electronics put in a human-like skin, making actuators move the way a human may - will confuse us even more. What questions or tests do we do if we have two human-like objects to differentiate which one is conscious?

Do we cut them open and see that one has organs and the other has cables? Do you cut every human today that you believe is conscious to prove that they are a human? Of course not (I hope).

Some would say that it's the internal world that living beings have that a machine doesn't. How do you know that a machine doesn't? A working processor is not idle, things are happening all the time - put an electrode or open the process manager and see how much stuff is happening. I am not an AI or machine learning expert, but I'm pretty sure a model can be made to evolve itself based on the user interaction. It can be let to explore whatever "it wants". And if the argument is - but it doesn't "want" anything, it's just instructions - what is the difference to a human brain "wanting" anything? Isn't it just instructions of the neurons directing us to do something, giving us the illusion that we want but it's actually predefined code? Where is the "wanting" Me inside of the brain? Where is the "wanting" It inside of the code?

We are conditioned to give a lot of value to a human brain and think it's very special. Energy-efficiency-wise - absolutely. We are magnitudes away from simulating a full brain even with all of our computers on the planet combined. We have a hard time simulating something that is drawing just tens of watts per hour. But just because nature over billions of years has blindly evolved something that is super efficient is not an argument that we can't do better if we put some engineering thought (for a few hundred or thousand years in some problems). As far as we know, there is no "magic" in the brain. And if there is, if we manage to discover it and measure it, it will no longer be "magic" just the way a rainbow appears magical but it's completely understood by us, we can make it, measure it, know when it appears and why.

So if we follow the assumption that there is no magic and it's just plain physics we know day-to-day, what happens if we simulate a brain in code. I'm still surprised to this day how many people consider that this thing is still not conscious. These people will generally agree that 1. a living brain is certainly conscious. Then they would also agree that 2. brain is just made of cells, which are in turn made of atoms, which are in turn understood how they work. But they would disagree that if this simulation is put on a super powerful computer and press "run", that "thing" will not be conscious. I can't follow this logic.

Of course this simulation argument also puts the question - what is the level of the simulation? Is each neuron a black box that we understand and try to mimic, ignoring all internal workings of the cell? Or do we go further down to organelles or even the atomic levels? If we go further down to the quantum levels, we might need a quantum computer as we will never be able to simulate anything. But does quantum mechanics really play a role in consciousness? Great minds like Sir Roger Penrose argue Yes. Great minds are "appealing to authority" so that doesn't mean much. Most scientists say No since the quantum effects are too small to propagate that many magnitudes. I have also written about this in the past. Or when simulating - do we go higher up in brain-regions levels and ignore minutia like the exact mechanics of neuron firing and abstract this away in layers of "knowledge" and linear algebra, which is the basic of how AI works today.

Well, this imitation of consciousness that the world is currently questioning is more of the latter - a rough simulation of the speech and memory regions, not of the emotional center or the motor center. But other efforts are going there as well, it's not like we can't put it all together.

We give too much credit to humans at this point. Brains and processors are following the same physics rules as far as we know. They have different chemistry and certainly different biology but so far we haven't discovered an element or particle responsible for consciousness. And all science is based on measurements - if you can't measure it, it's a belief system.

So after all of that what is my belief - is a computer conscious or not? I would answer the same way if asked in 1800 if a material is radioactive or not, in 1600 if the dots on the sky have attractive forces like the Earth or if the rainbow is magic - "humanity doesn't know yet".

And how about the claim of a certain engineer that a program is conscious? Well, humanity has literally "fired" people for wrong religious beliefs in the past. The current accepted dogma seems to be that humans are definitely conscious (for sure from the moment of birth, some would argue before that) and various animals might be considered conscious. But unless you are a fringe philosopher or certain substance abuser, a rock is certainly not conscious. So going against the dogma will lead to "firing", at least today it's in a much softer "burn".