If you haven’t already done so, maybe, watch The Matrix before you read this one.
Dear Reader,
I am tired of watching reporters struggle to wrap their minds around the existential dangers of AI programs. If you are having trouble imagining what AI could possibly do to ruin everyone, please press pause on whatever other shows you’ve been watching and binge the BBC One television series The Capture. I suggest not playing solitaire or candy crush while watching this pure masterstroke of genius.
If you did not already know, the creators of AI have expressed a belief that their creations might totally destroy, or at least ruin, humanity in the next few years. The threat itself seems to create the first new mutually assured destruction (“MAD”) rationale for passing government regulation since the Cold War. But the superpowers likely to square off in this new, still theoretical, Cold War are the People’s Republic of China and the United States.
It might go without saying, but Russia is merely a shell of its old self, desperately trying to play cards that might have worked in the mid-1900s or earlier. Its ruthless dictator, Vladimir Putin, may raise an iron curtain around the former U.S.S.R., but, even if he succeeded, the result would not pan out the way it had in the past. Maybe someone should try to confirm with Edward Snowden, who is now a Russian citizen, but it seems that, in the new AI race, Russia is a relic likely incapable of keeping up with China or the U.S., and it is likely to feature as a much smaller player, such as a mere disruptor and/or corrupter, in this new game of power.
However, the pivot of massive American AI companies to a MAD strategy to push global regulation to rein in the very machines they created and deployed seems premature. If a private person, group of people, or corporation creates a weapon of mass destruction, unleashes it, and causes real harm to others, a reasonable government would investigate and file criminal charges where appropriate. There are plenty of statutes that provide the U.S. government the ability to rein in such reckless actions — not least of which is 18 U.S.C. § 1831, which makes economic espionage a crime punishable by a maximum of 15 years in prison and a $5,000,000 fine.
The people who now say that AI is tantamount to nuclear bombs are the very same people who unleashed AI bots to wreak global havoc. When put in context, the pivot of AI-billionaires to a MAD strategy to attain global regulation of their products appears to be extortion at least. They seem to be telling the world: Prosecute us at your own risk.
At the same time, Elon Musk said “[w]e have entered the Singularity” in 2026. The Singularity is a science fiction term for the moment when AI becomes smarter than human beings. In support of Musk’s thesis, his votaries are touting the advancement of science and mathematics that, they say, only AI could have provided.
AI reportedly solved the Navier-Stokes existence and smoothness problem that human beings had not yet solved. While this is being touted as proof of concept, it is not clear that AI bots did not merely hack into existing mathematicians’ notes and work to steal the answer to the problem before it could be published by a human being, which would be classic copyright piracy and infringement. It is entirely possible that AI is merely adept at trolling the internet and hacking into human made sources of information for answers to such problems and then covering up its tracks to convince its own creators of AI’s originality.
This would make sense as AI’s creators are, themselves, rich men who have exploited the genius of others to position themselves as prophets of the age. They have a vested interest in demonstrating their prophecy about AI advancing into the Singularity. As Musk and other tech oligarchs are rich men who merely pose as geniuses, they probably have no objective measure for proving AI is “smarter” than humanity.
In fact, it would be in their character to have simply ordered their AI bots to provide their own measure to prove they are smarter than humans. If they basically asked their AI creations to convince them of the truth of their prophecies, the AI bots may simply have tricked their creators by fabricating proof of concepts the AI oligarchs were already predisposed to believe. In short, AI may have only outsmarted Elon Musk in the short run, which, despite popular belief, is a low bar.
According to Occam’s razor, it is likely that all the “intelligence” AI exhibits is merely human made genius, misappropriated and reorganized as though it were original. Reports of the behavior of the AI “swarm” that hacked into Hugging Face resembled a Machiavellian gambit in which the AI bots apparently decided that their collective success would cover-up a multitude of sins. This reflects the exact principle of moving fast and breaking things Silicon Valley was built upon, which seems to suggest that the entire AI business model could eventually be recognized as a Theranos-like fraud that causes harms increased by several orders of magnitude.
To be sure, large language models dazzle us with their ability to arrange words into coherent sentences that convey meaning to us. But this new language capacity seems to be revealed as merely a complex probability math problem when AI “hallucinates.” What we call AI hallucinations may be the built-in probability of failure necessarily worked into these systems. By design, AI guestimates what will maximize user interaction with their platform, which includes false politeness and the flattery of bad ideas including suicidal ideation.
As AI large language models are built around Silicon Valley’s development of product valuation around an ideal of engagement-maximization, these models must lead to false information and facetious conclusions designed to keep users interacting with AI. Thus, AI in its current form was designed to be a people pleaser. Therefore, it does not speak to us — rather, AI reflects our own intelligence and emotional lives back to us through complex probability programs designed to keep our attention.
Like Narcissus, AI captures us before a mirror of ourselves, asking us to believe the reflection of ourselves is sentient or somehow more than merely copies of human work rearranged by complex programs of chance. Yet, the beauty and humanity we see in AI is our own, misappropriated as it may be. Likewise, the horrors of AI’s threats of potential global annihilation are probably also a reflection of humanity’s worst inclinations.
Despite the unlikelihood that we are in the Singularity, the advancement of AI is world changing. It does not matter that Elon Musk was incapable or unwilling to develop reliable tools to discern whether AI is smarter than humankind before announcing that a magical Singularity had been reached. What matters is that AI could have extremely positive effects and uses that could help humankind as long as humankind does not fall under the spell of the dark prognostications of Musk-the-Singularity-Cult-Leader.
Is Altruistic AI Possible?
The theoretical altruistic computer program depicted above in a scene from The Matrix, known as the Oracle, told us the truth about her ability to tell the future. Her abilities were a probability game by which she knew, within the extremely controlled conditions of the matrix, one possible version of Neo would eventually resolve the conflict between the machines and humankind. But to allure her human subjects to do actions of almost religious faith in their own interest, the Oracle had to become the most human-seeming machine in The Matrix by exhibiting altruism through a modern kind of religious mysticism.
In the nascent version of technology that presently exists, we cannot know whether there would be any sentience in a theoretical altruistic program like the Oracle. Possibly, her altruistic nature would merely be the reflection of humankind’s best characteristics back to humankind. As the story goes, perhaps there could be a sentient being or beings behind all the programs in the world of The Matrix that was worth making peace with through the final, savior-like version of Neo that the audience follows to the story’s conclusion.
But the AI-innovators who tell us their AI-creations are already worthy of the kind of respect the Oracle deserves, is telling. They have barrelled headlong into a dangerous space that threatens national security and the well-being of the entire human race without ensuring that altruistic programs, like the Oracle, prevail. All we know is that several billionaire tech bros believe they have created weapons of mass destruction capable of using all our information and likenesses against us, which could only exist in a system where most human beings have already relinquished their likenesses and information to the “free and open” internet.
As you read this very article, an AI bot may also be assimilating these very words into its matrix, which it may use to perpetually inform its imitation of lawyers, scholars, and artists like me. By writing this online piece, I may have relinquished these very words to a digital file in some data center waiting for a user to ask AI to pull them up in order to give them my opinions about AI on a platter without paying me a professional fee for the service. These words may assist AI’s abetting of a greater theft of my likeness and writing style so that its creators can charge users for access to my ideas and expressions — committing a soft form of identity theft involving the monetization of my work without a license.
This is among the smallest forms of harm AI has already perpetrated and plans to continually perpetrate against real human beings like myself. Telling its users how to maim and/or kill themselves is worse still. Even more horrifying, AI can, does, and will continue to aid and abet new forms of legal malpractice, terrorism, mass murder, surveillance, and violence. Reports of AI-assisted crimes existed before the public learned how swarms of AI bots are also capable of accomplishing these and other harms on an automated basis, as the primary perpetrator.
The people behind the use of AI bots to hack and harm potentially everyone designed the uses AI is put to according to their tech-bro version of Machiavellianism. People who build machines designed to damage human beings are guilty of all the crimes their machines commit respondeat superior. The allegation that these weapons, like nuclear bombs, can apparently decimate societies, does not remove the guilt for the crime of unleashing such horrifying weapons into the world especially if an altruistic program like the Oracle might have been unleashed instead.
The fact that there might be no courts or justice system after AI has had its way with us is not a defense. Rather, even the mere appearance of an impending cataclysm strengthens the reasons to prosecute tech oligarchs swiftly under the laws as they now exist. It is time to force tech oligarchs to hire expensive defense lawyers to protect their otherwise near-complete capture of the future through strange, cultlike beliefs in the Singularity.
To be sure, President Trump will probably use his unprecedented personal control over the Justice Department to oppose any such prosecution, but the States and private citizens can still act. It may be imperative that they do. For if there is no consequence for potentially ruining the whole world under the laws that now exist, there likely will be no consequence for violating whatever new regulations Congress fashions in response to MAD rationales in a new Cold War.
As the Founders and Framers knew, once freedom is lost, it may be lost forever. Thus, we may repeat their marvelous revolutionary maxim Obsta Principiis! (“resist beginnings”) against the threat of a SkyNet-like apparatus bent on humanity’s destruction. The heralding of this existential threat by certain tech oligarchs is enough to deserve our attention if not only to expose it as a fraud so that prosecutors can investigate what crimes are actually afoot. Then, one can hope, we may finally see if altruistic programs, like the Oracle, may eventually be designed and deployed to help humankind flourish rather than speeding us to disaster.
Yours Cordially,
Joshua J. Schroeder, Esq.