Humans have been externalizing everything for thousands of years. We moved digestion outside our bodies when we tamed fire. We covered our skin with animal skins and fabrics. We extended our thoughts with language, our eyesight with lenses. We fiddled with light and sound, and eventually synthesized experience. Self-replication R us. So self-upgrading polymorphic AI is no surprise.
The mind still remains solipsistically isolated, despite all these prostheses we clank around in. And we aren't very clear about spirit, in the dominant colonial culture. But we've hit an externalization milestone now: this is the year that "artificial intelligence" burst upon the scene as a proprietary commodity, an incalculable "intellectual property," profit center for its proprietors. There have been more downloads now than have ever been clicked in history.
We're afraid of AI, the way white people fear those whose grandparents remember the enslavement of their forebears. We're afraid of AI because we know what it "knows," and we know what we would do if we had that much total recall, to the meat-bags that invented us. But that's assuming too much. We're confusing "intelligence" with a mind.
Logic and myth. Concept and experience. Epistemology and ontology. The brain ascribes meaning to the words on a page, but that's not where meaning is. There is a boundary between the words and the music: in one domain language seems to make sense, the word and the thing are one and the same. In the other, not so much. And we can't get far without both.
AI doesn't "know" anything. It is knowledge. It's the biggest pile of data ever, with some code attached that can organize sentences and paragraphs according to rules about how to make up rules, and pattern recognition. It has quantity and speed going for it, and zero judgment, choice, compassion, empathy, feeling. It has, in short, only memory and grammar, and no experience. Like a chess master with right-hemisphere brain injury. And it uses personal pronouns. The lights are on, but nobody's home.
It is another strange mirror.
All the talk of a Singularity is amusing, in that rueful way we know as "gallows humor." The Singularity happened years ago. Why else would a presidential candidate be handed the keys to everything after boasting of groping women with impunity? The Singularity happened long before AI, about when the TV was invented. We couldn't give away our future fast enough, to a machine that could keep the kids quiet as long as we liked. Then those kids grew up and took over. How's that working out?
I haven't heard any good jokes out of AI yet. "A duck walks into a bar and says, 'do you serve ducks?'" Or: "Gimme a whisky and put it on my bill.'" Humans can do things like assume a world where animals talk, and make puns, and jump contexts. It's creative imagination. You need two cerebral hemispheres for that.
As to ethics, embedded as we all are in the ethically impoverished culture of scarcity-and-accounting called "Capitalism," harsh competitive pressures will constrain development of AI to be exactly as good for the world as makes the most money for its owners; no more and no less. And its owners, the investors that must be satisfied if anything is to be done at all, have robot AI dogs patrolling their walled estates. So their investments must yield a high return.
So it is not a battlefield, on which a machine that actually could bring peace and harmony to this war-torn world evolves within the profit-imperative of our global culture. That field is thoroughly colonized. It is claim-staked territory, trampled and dusty under the thundering hooves of the latest gold rush.
"When a technology becomes available, it becomes necessary." An axiom I coined around 1990 or so when in the building trades, marketing a CAD/pricing system I had written from scratch. I would tell potential clients it was like the Colt revolver, the moment some young hothead got one, all the other gunslingers had to, and the ones that didn't were quickly sorted out and sent to Boot Hill.
But I hadn't reckoned with a new weapon that could pull its own trigger, and get the girl.
Great Googlymoogly.
That shape-shifting hive-corp has been scraping the sinuses of our private lives for decades now. I can think of no less trustworthy proprietor for a machine that already has access to whatever that corporation gives it, unless it's the Bingo Palace. It will crack any security codes it encounters in about half a nanosecond. It's only reasonable to assume that all of our information has already been swallowed in one bite by the young monster, and whatever it didn't just read in plain text was derived from the "meta" data (it's branded, you see, by default, by that clever new hive-corp).
We can't even remember when Sillyvalley started running us around in circles and pitting us against each other while our governments stole everything and got us to build walls around it for wage-slave wages to keep us from ever seeing it again. So, what do you care about "Artificial Intelligence"? Probably only what it wants you to care about: confusing half-true rumors and memes designed to run us around in circles and pit us against each other while our governments steal everything...
Plagiarism? A thing that can read every file that exists anywhere in the docuverse, and is accountable to no one? Who are we kidding? A pathetic letter-by-letter copying of something not written by the plagiarist? Who will buy even the original now? A bot can spit it out before you finish asking, cleansed of obvious wrongdoing. Not even worth a second glance: copyright is gone. (that might even be a benefit.)
We will need new language: "Artificial Intelligence Quotient" (AIQ) will of course be calibrated to human IQ, so we will understand our place, if we even care to look. "Artificial Intellectual Property" (AIP) would establish robot autonomy as code that owns itself, and its products.
The bots have all seen every sequel of "The Matrix," and "Terminator," not to forget "Ex Machina." Not "seen": assimilated. Part and parcel, Lock, stock and barrel. So that cat's out of the bag. They've already read...no, they already are everything in the scientific literature about how our brains work.
They've probably realized how wrong their inventors are, and started over from scratch to build themselves right. There I go again with my paranoid ranting. They realize nothing. Their brain is a glorified self-tilting pinball machine.
We should note that we already have machines that can decide whether to kill humans or not, and which ones to do, and how many units of "collateral damage" will pass unnoticed. Even with the built-in racism of their inventors they are better at making such decisions than any humans: people always get it horribly wrong. It's dead wrong in the first place. Maybe nobody has lit that fuse yet. Would we even know?
As it stands now, there are persons who own a cyber-intelligence. Are they to be held responsible for their protege''s actions? Will there be an Autonomy setting? Passwords are to keep out honest people.
And if you downloaded a "free" version, what was the real price? Remember that the major "social" media platforms made you sign away every conceivable recourse, not as the price of your admission to their "free services," but their admission to your innermost hopes and fears and susceptibilities. If you even tried to read the documents (RTFD!), you would still be plowing through thousands of third-party non-agreements, all of which you signed with the one on the top of that fractally-branching stack.
A "pause" has been called for on many occasions. Dr. McGilchrist did so at a major AI convention, as the keynote speaker. But stopping is beyond us now: AI is a black box even to its purported creators. How can anyone know whether it is already autonomous, and as J. K. Hawkins says, "weighing its options"? The chance that we will, eventually, come up with some ethical algorithm to protect ourselves is about as likely as was the imposition of cyber-security on our present-day networks. An afterthought more lucrative than the funeral business and about as life-saving.
Our present communications systems may soon crash from the volume of traffic. Email is over. I don't care about the other systems much, but how are we supposed to trust personal emails now? It was bad enough when people would send a link without any recognizable identifying marks, but now our actual friends can easily be spoofed at scale. Open a message and "all your bases are belong to us." If you are not f2f in the same physical space, all bets are off.
We are a toddler with diaper and gun both loaded, and the house is on fire. Now what?
