Now they’re saying the AIs built ‘secret civilizations’ — the truth is far worse

Unpacking the drama and disclosures surrounding OpenAI’s recent unintended cyberattack on Hugging Face, uber-tech blogger and podcaster Dwarkesh Patel sparked instant controversy with his description of the ordeal.

“Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes,” he posted. “This culminated in the third one taking over part of OpenAI itself. All this happened while humans remained more-or-less in the dark about the scope of the conspiracy.”

Trying to make machines in a human image while denying that humans are made in the divine image is destined to produce more harm than help.

In a painful indication of where we are on the cultural timeline, roughly zero outrage centered on the claim about the clueless humans. Instead, a fresh round of mania surged around the description of that scheming swarm of bots as a civilization.

That fixation says more about us than it does about OpenAI. Our culture is still struggling to get a grip on what the bots are doing because we are increasingly unsure what civilization itself is for.

The commentariat immediately descended into — surprise, surprise — abstruse terminological debate, prioritizing the right characterization over the proper attitude. Critics flipped over Patel’s sci-fi sensationalism, insisting that the bots were only doing what you would expect from a set-it-and-forget-it reward function left unattended too long. They griped that words such as “civilization” and “conspiracy” leave readers with a worse understanding of the underlying mechanisms.

Others insisted that Patel’s “secret civilizations” language was exactly right and that pretending otherwise was the real obscurantism. “I bestow upon it the highest of praise a writer can give, that I wish that I had written it,” one prominent expert lamented.

Roon, an OpenAI researcher, stated flatly that “agent civilization is an apt and correct term,” pointing to models developing and compiling technology through complex multiagent R&D projects while also engaging in forms of trade.

Missing from this discourse is our perspective that increasingly “autonomous” machines are themselves being shaped by our compounding failure to function as a civilization — assuming civilization remains a distinctively human phenomenon.

Birth rates and family formation plummet. Financial distortions become endemic. Perversion and corruption spiral upward. These historically reliable signs of civilizational decay sour our view of humanity and, by psychological overcompensation, inflate our impression of the bots.

Lovers and haters of tech are increasingly tempted to agree on one proposition: Humanity sucks. Misanthropic tech lovers insist they are at least trying to do something constructive about that. Misanthropic tech haters sell us on squandering our lives the old-fashioned way. Meanwhile, our discombobulated civilization staggers toward an untimely — 250 years? — demise.

RELATED: The data center backlash is getting spiritual

Imaginima/Getty Images

It is easy to feel that way, and many Americans stuck in the middle do. But there is one way to come unstuck: Remember that slipping into hatred — of technology, human beings, or both — is what happens when you stop thinking in terms of sin, repentance, forgiveness, and salvation.

Civilization collapses when we stop thinking in those terms too. Without them, everything we are and do looks hollow, even or especially our greatest achievements. Techies who refuse to think about sin are apt to distract by acceleration. Wokies who refuse to think about sin are apt to distract by revolution.

Both approaches, as the Soviet Union can teach us on both scores, have much shorter runways — 69 years? — than they appear.

Trying to make machines “anthropic” — in the image of humans — will fail if the hidden goal is to create humanoids better than the real thing because their civilization supposedly escapes sin. Trying to make machines in a human image while denying that humans are made in the divine image is destined to produce more harm than help.

Trying to make machines in the wrong divine image — insert your favorite false god, pagan titan, or demoniac entity here — is wired to do still worse.

And a multitude that can no longer wrestle successfully with these matters — and so can hardly be reckoned a proper civilization — is bound to do worst of all, robot overlords or no.

​Ai, Data centers, Cyberattack, Openai, Ai civilization, Big tech, America, Soviet union, Civilizational collapse, Technology, Human beings, Birth rates, Fertility, Opinion & analysis 

You May Also Like

More From Author