John Rachal's Blog

“People will come to love their oppression, to adore the technologies that undo their capacity to think.” Aldous Huxley, Brave New World Revisited, 1958, discussing themes of his 1932 novel Brave New World

We are building the apocalypse, and we are building it with more siren seductions than there are stars in the heavens. Artificial Intelligence will infinitely ease and expand our lives—that’s the gist from its developers and apologists. In some ways, perhaps; but at a potentially existential cost. That apocalypse won’t happen in my lifetime, or probably my daughter’s, or hopefully not my grandchildren’s. Perhaps—let us hope—I’m just Jeremiah, railing not at things as they are but at a dystopian technological future, never to be fulfilled.

Yet it is indisputable that this technology is profoundly dangerous. Possibly as we keep playing our pungi, the cobra will go back into the basket. But I don’t think so. Students having ChatGPT write their term papers, or professors sending fraudulent AI-generated manuscripts with fake AI-generated data to journals could be the least of our problems. One survey cited in Time found that almost half of AI researchers agreed that there was at least a 10% chance that AI could exterminate humanity, presumably at some distant future point. In the 1968 movie 2001: A Space Odyssey (over half a century ago!), the onboard computer HAL has become effectively sentient and attempts to take over the spacecraft, and in 1984’s Terminator the system Skynet, at a designated future date, will become “self aware” and have even more sinister ambitions.

Hollywood fantasies, of course. But is it impossible to think that we could blunder our way into creating a technology dictatorship that no longer serves us (for good and ill) but competes with us, and ultimately considers it in its best interest to destroy us? And all along the way to expand the opportunities for human actors to use these technologies in nefarious ways, including committing political violence, or even starting wars? We already know well the potential of social media and bots to stoke fear and anger and spread disinformation creating hatred and havoc. Shockingly realistic audiovisual deepfakes are already here. Facebook and QAnon effectively killed people during covid. We have seen Russian technological capability interfere in our elections. January 6^th could not have happened without the internet. These are merely the warning shots of a very dangerous future, a new and very dangerous brave new world.

As Charlie Warzel of The Atlantic has noted, others have drawn comparisons between Artificial Generative Intelligence and other paradigm-shifting technologies, such as the internet itself. But he notes that those others do not make a comparison between AGI and the advent of the nuclear age. He might have added that when Robert Oppenheimer, the father of the atomic bomb, witnessed the first atomic detonation near Los Alamos in 1945, he soberly quoted a line from the Hindu Bhagavad-Gita: “Now I am become Death, the Destroyer of Worlds.” Humans have now managed to go seventy-eight years since Nagasaki without using an atomic weapon in anger. On the plus side, nuclear energy—a necessity in the fight against climate change—has been a clear if worrisome benefit, just as AGI has benefits. Yet the danger of nuclear military catastrophe, of course, is still there, though so far so good. Will we be as lucky with Artificial Generative Intelligence? And for how long?