“AI” enthusiasts arguing for “AI” consciousness say: prove to me that you are conscious. Of course, one cannot prove this. So this is taken as some kind of “gotcha,” proof of an unfair double standard. The enthusiasts then suggests that this means “AI” consciousness is just as plausible as human consciousness.

Here I propose an explanation for why we can know that other people are conscious, even if we cannot formally “prove” it.

The argument is simple: if you are conscious, and you know that other people have similar behaviors and biology (especially brains) to you, then surely they must have similar consciousness to you. The alternatives are too implausible to take seriously.

Likewise, we can know that animals have some consciousness: the more similar that animals are to us, biologically and behaviorally, the more-similar their consciousness ought to be as well.

I am not proposing a solution to “the hard problem” of consciousness, meaning I do not offer an objective test of whether any “AI” or alien is conscious, nor do I offer an explicit definition of consciousness. My arguments here do not disprove the possibility of “AI” consciousness. However, they highlight the need to consider mechanisms (biology and/or algorithms) and not just behaviors when attempting to judge consciousness.

The problem

You, the person reading this, are conscious. You have some feelings, experiences, and awarenesses of each moment of your waking life. How do you know you’re not the only one?

Your behaviors relate to your inner life, for example, acting differently based on how you feel, and telling other people about it. And, for your whole life, you’ve observed other people talking about their inner lives, and behaving in ways that seem appropriate. You act one way when mad, and another when relaxed. You can recognize the relationship between others’ behaviors and their stated inner lives, based partly on your own inner experience.

But perhaps it is an illusion, or a lie. Perhaps everyone else around you is basically a machine, performing the behaviors as though they are conscious people like you, but they are not. Perhaps they are all “philosophical zombies.”

Two sources of information

We have two sources of information about other people: their behaviors, and what we know of their biology.

You and I are both human, part of the same species and the same evolution. We share, along with the rest of humanity, the same kinds of brains, hearts, nervous systems, and so on. We can eat the same kinds of food, we need the same kind of exercise and sleep and so on. Of course we mostly focus on the differences between us (say, food allergies, different tastes or preferences, neurodivergence), but it’s not like some of us can hibernate for six months at a time and others can metabolize gravel.

Our brains mostly have the same parts and the same functional elements at different scales. So it makes sense to believe that our mental states are similar too—we share the same broad palette of emotions and mental representations, even though they are expressed differently in different people.

Hence, these facts together give you every reason to believe that other people are conscious:

You are conscious
Other people are biologically very similar to you
Other people talk and behave in ways consistent with your experience of how your behaviors relate to consciousness

Indeed, we naturally take others’ behavior as cues to their mental state, and might question the mental wellbeing of someone who’s behavior seems inconsistent with a sensible mental state.

This, here, is sufficient evidence to live your life without seriously doubting other peoples’ consciousness. Doubting other peoples’ consciousness is purely an intellectual excerise, not something worth taking seriously in how we understand the real world.

The alternatives: paranoid conspiracy theory or irrelevance of consciousness

After all, what is the alternative?

It could be that you alone are conscious, and everyone else is merely performing behaviors of consciousness, despite sharing the same biology. This seems to me a paranoid conspiracy theory, like imagining that we live in a simulation or some other constructed reality. No one can disprove such fantasies, but there’s no genuine reason to believe them either, and, even if true, they’re irrelevant to any decision we’d make in our lives.

It could also be that most people are conscious, but some rare individuals lack consciousness while still functioning in society. There are lots of examples of rare individuals that lack other perceptual or mental attributes, such as colorblindness, aphantasia (lack of mental imagery), stereoblindness (lack of stereo vision), to sociopathy (lacking compassion for other people). Some of these are hard to detect, but they are detectable (even though sociopaths might perform compassion). If there were philosophical zombies amongst us, you might think that we would have already discovered it somehow.

If there were no genuine differences in behavior between people with consciousness, and philosophical zombies (who claim to experience consciousness but do not), then that seems to imply that consciousness is irrelevant for normal life. That is, consciousness has no noticeable impact on a person’s behavior or ability to function; it just randomly appears in some individuals but not others.

Each of these alternatives seem so implausible as to be unworthy of serious consideration.

The only sensible possibility is that everyone who is awake is conscious.

The varieties of consciousness: altered states and animals

How can we know that animals are consciousness?

Note that it’s unlikely that animal consciousness is identical to human consciousness. Perhaps animal consciousness is like human consciousness, but with some differences. Perhaps, day-to-day, my dog experience a vivid alertness that’s a thousand times more than my own peak (like when playing a vigorous sport or video game), but lacking any mental foresight (like when I’m drunk and groggy and barely able to function). There’s reason to believe that octopuses are conscious in very different ways from humans.

Regardless, the same reasoning as above can tell us that mammals are conscious. Consider chimpanzees and dogs, each of which are social animals and each is related to us biologically, including in terms of brain structure. They each exhibit behaviors that seem to indicate consciousness, such as social expressions of inner states (happiness or anger, etc.), and other sorts of behaviors (attention, distraction, etc.). We share similar internal organs, similar brains as monkeys and dogs, and some of the same behaviors. We share evolutionary lineages with them as well; they are evolutionary cousins. Hence, they ought to share some sort of consciousness as well, by the same reasoning as above.

And, the alternatives to animal consciousness seem highly implausible. If no animals are conscious, then consciousness emerged solely in humans. But nearly all of our faculties evolved in earlier creatures; monkeys have many of the same faculties of intelligence as us, as do their ancestors, and their ancestors. Consciousness seems most closely tied to mood, emotion, alertness, awareness, social behaviors—all attributes shared by many animals—and not to the few things that seem to make us biologically unique as humans, such as the ability for complex spoken languages.

It seems likely that more-sophisticated intelligence goes along with more-sophisticated consciousness. Dogs have much less-sophisticated intelligence than humans, and so their consciousness would be much less-sophisticated as well, but still present.

Machine consciousness and alien consciousness

We cannot apply the above arguments for machine intelligence, because machines do not share biology with us. Current “AI” algorithms do not at all operate like human brains. There are ways in which they do operate similarly, e.g., convolutional network models of the visual cortex, dopamine as modeled well by reinforcement learning, but these are small pieces of the overall functioning.

This does not disprove “AI” consciousness, but it shows that establishing “AI” consciousness is far, far, more difficult than human consciousness. Which seems obvious, but apparently is not to some people.

How would we demonstrate or prove machine consciousness? Or, if aliens came down from space, how would we determine if they are conscious? I don’t think anyone has a good answer. However, I do not believe that we can judge consciousness simply by looking at behaviors, like performing Turing Tests. We’ve already seen over and over that people can be fooled by simple rule-following chatbots like ELIZA. We can build convincing philosophical zombies.

How do we know that chatbots like ELIZA are not conscious? We read the paper, or look at the code to see how it works. We can know enough about how LLMs work that to know that they, too, are simple text processing engines that, like ELIZA, literally do nothing more than take text strings as input and generate text strings as output.

It might be that computer algorithms could be described as “conscious” if they have any mental state representation at all, like state variables representing “mood.” But such representations are so different from human consciousness that calling them “consciousness” in public discussion seems misleading and extremely irresponsible (a topic I intend to exapnd on in a future blog post).

All of these arguments indicate that judging consciousness, we have to consider the underlying mechanisms, that is, the algorithms and/or the biology, and not just inputs and outputs. Similarity to our own biology would be a positive clue; algorithms that generate print statements as a functions of inputs would be a negative clue.

Summary

I have argued that you can know that other people and animals are conscious, because you are conscious, and because you share important behaviors, biology, and evolution with other people and with animals. This is not a solution to the “hard problem,” and it is not rock-solid scientific proof. But you can be confident in it enough to assume that it’s true—the alternative involves paranoid conspiracy theories or philosophical zombies well-hidden among us. On the other hand, there’s no good reason to believe that algorithms, which have totally different development, substrate, and behaviors from us share this one hidden attribute.