AI as philosophers' stone?

As American civic life has become increasingly shaped by algorithms, trust in government has plummeted. Is there any turning back?

"...In “The Techno-Optimist Manifesto,” posted last October, Marc Andreessen, the venture capitalist and notable Trump-Vance supporter, delivered a delusional account of human history as the triumph of the “techno-capital machine” over the constraints of nature. “We had a problem of isolation, so we invented the Internet,” Andreessen proclaimed, preposterously. The solution is at hand: “We believe Artificial Intelligence is our alchemy, our Philosopher’s Stone—we are literally making sand think.” This September, OpenAI’s Sam Altman posted an essay in which he argued that generative A.I. is “the most consequential fact about all of history so far,” and that humanity is on the cusp of solving every problem. “There are a lot of details we still have to figure out,” Altman wrote, unironically, “but it’s a mistake to get distracted by any particular challenge.” The following month, not to be outdone, Dario Amodei, the C.E.O. of Anthropic, a rival of OpenAI, published a blog post called “Machines of Loving Grace: How AI Could Transform the World for the Better,” in which he predicted that A.I. could lead, in five to ten years, to “the defeat of most diseases, the growth in biological and cognitive freedom, the lifting of billions of people out of poverty to share in the new technologies, a renaissance of liberal democracy and human rights”—developments that will happen so fast and be so overwhelming that many of us will be “literally moved to tears.” Someone will be crying. That much is true.

Having built an information infrastructure that classifies and divides humans and drives them to ideological extremes, these same people and corporations are now building machines that purport to undo the very damage they have caused, much in the same way that geoengineering schemes seek to address catastrophic climate by using the very logic and tools that created the problem. In a study funded in part by M.I.T.’s Generative AI Initiative and published in Science this fall, conspiracy-minded Americans were subjected to long exchanges with a deprogramming chatbot. “The treatment reduced participants’ belief in their chosen conspiracy theory by 20% on average,” the researchers concluded. They don’t seem to have bothered to establish control groups who might, for instance, have been asked to read articles and books, or—seemingly beyond the realm of imagination—converse with another human..." Jill Lepore