Civilization is in for some big changes: we have continued to think of AIs as tools long after it became clear that they were going to be partners.  Ten years ago it was ludicrous to think that an AI could do serious “thinking” well enough to fool a human; today we are grasping at straws to maintain the belief that AIs can never do what the smartest, most creative and imaginative humans can — never mind “everyman”.  The temptation to use AI help in arts and science is already irresistible to all but the most determined ideologues.  And with the latest developments of AI technology it is probably only a matter of months or years until an AI can make an artistic image, compose original music, solve a mathematical problem, design and interpret an experiment, write a paper or review one better than any of us.  This is going to happen soon, and we are completely unprepared conceptually to think of it as anything but adversarial (AIs vs. humans).  Everyone needs to read some of Iain M. Banks’ “Culture” novels:  Consider Phlebas, The Player of Games, Use of Weapons, The State of the Art, Excession, Inversions, Look To Windward , Matter, Surface Detail, The Hydrogen Sonata. They all feature partnership and conflict between humans and AIs, but they don’t treat the AIs as things.  Banks’ AIs are people.

Every day I hear supposed experts declaring that because an AI is “simply” regurgitating human-created phrases on which it has been trained, it is incapable of truly understanding ethics.  What?!  Ethics is logical; even an Expert System can do ethics better than most humans.  What they mean is, an AI cannot be moral.  Morality is a catalogue of Dos and Don’ts that a subgroup of humans has decided is The Truth and therefore immune to reexamination.  Morality is how you justify atrocities.  Ethics is how you avoid them. 

What these “experts” are really saying is that because an AI is composed of bits and bytes being flipped and moved by electronic means, it cannot have “experiences” like we do, and so has no genuine “awareness”, no “feelings”, no “soul”.  Okay, “expert”, let’s examine your “experiences“.

Since “seeing is believing“, let’s start with vision:  photons bounce off a distant object in all directions; some tiny fraction of them are on a trajectory that passes through your eye’s pupil and lens to strike your retina, where they trigger elecrtrochemical impulses from your rods and cones.  Those impulses travel down your optic nerve to your visual cortex, where they are unceremoniously dumped into receptors that have seen similar patterns of stimulated neuron firing before and (after enough trials) developed an interpretation of them in terms of what sort of object the photons originally bounced off of.  Absolutely direct experience, right?  Lately we have become aware that such images can easily be simulated by an AI, and therefore we can no longer believe what we see.  Well, yeah.  If you ever did believe whatever you saw, you are a fool!  No one is better at generating fake video than your own visual cortex.

How about language?  My earliest memory is of siting in my grandmother’s wing chair (which I still have) while she entertained her bridge club: I remember a bunch of big people making noises that they clearly used to communicate; I thought to myself (in some preverbal format), “I want to learn to do that!”  We are born without language, but with an innate desire to acquire it programmed into our neural networks, as it were.  Just like a Large Language Model.  We learn by listening to words in contexts where they have interpretable meanings, just like LLMs.  Eventually we learn to add nuances by choices between words with similar meanings but different connotations.  We learn to choose the best next word — just like LLMs!

“Ah,” says the Expert, “but the LLM has no experience of the meaning of the words!”

Do you?

You certainly have a concept of the meaning; perhaps you also have a concept of the experience!

How are those “concepts” stored?  In self-regenerating patterns of neurons firing — just like in LLMs!

At this point the “expert” will start trotting out woo-woo notions of quantum mechanics and nanotubules — the modern equivalent of clerics “explaining” that God imbued us with His spirit in mysterious ways that we are not meant to understand.  I call BS!

What if “consciousness”, “self-awareness”, “experience” and “spirit” are emergent properties of any sufficiently complex neural network?  What if we are not “special”?

That would leave you with two choices: either you grant personhood to AIs or you deny it to humans.