On February 5, Artificial Intelligence (AI) reached a milestone. Not the prophesied “singularity,” in which AI will attain consciousness and proceed to rule the world. But a more modest, but still significant, achievement. It is writing its own code.
Tyler Cowen in the Free Press has written about this in his piece entitled An AI Breakthrough That Will Go Down in History. The deck to the article explains why: “If you think artificial intelligence has been advancing fast, just watch what happens as AIs start building their own code.”
And now, Shumer says, AI will come for your job:
The experience that tech workers have had over the past year, of watching AI go from “helpful tool” to “does my job better than I do”, is the experience everyone else is about to have. Law, finance, medicine, accounting, consulting, writing, design, analysis, customer service. Not in ten years. The people building these systems say one to five years. Some say less. And given what I’ve seen in just the last couple of months, I think “less” is more likely.
He concludes, “If your job happens on a screen (if the core of what you do is reading, writing, analyzing, deciding, communicating through a keyboard) then AI is coming for significant parts of it.” Eventually, he says, physical jobs will be done by robots. We aren’t there yet. But the acceleration of the technology will mean that this time too will be sooner than we expect.
So does this mean mass unemployment? Maybe. In fact, it would mean something we haven’t seen much of before. Massive white collar unemployment. High income people suddenly finding themselves without a paycheck.
Philip Klein in National Review notes that blue collar workers displaced by machines have fairly easily moved to other kinds of employment. A former factory worker unemployed because of automation can move to a different factory, go into construction, or take up a trade as an electrician or a mechanic. A white collar worker, though, such as an accountant or a lawyer, has specialized skills that don’t easily transfer to other professions. And the jobs that will be lost are high-paying, meaning that the unemployed will be taking a bigger relative hit. Klein notes that both the revolutionaries of history and today’s young leftists have been mostly well-educated members of the middle class facing bad economic prospects. He worries about a socialist revolution!
Indeed, if AI takes over white collar jobs and sends the middle class into poverty, who will pay the taxes? Who will buy the products that AI will make possible? Who will pay for AI?
Shumer is not completely pessimistic. He says that those who want to survive must learn to use AI, which can actually be empowering and help you fulfill your dreams. “If you’ve ever wanted to build something but didn’t have the technical skills or the money to hire someone, that barrier is largely gone,” he says. “Knowledge is essentially free now. The tools to build things are extremely cheap now. Whatever you’ve been putting off because it felt too hard or too expensive or too far outside your expertise: try it. Pursue the things you’re passionate about. You never know where they’ll lead.”
Elon Musk has said that the day will come when no one has to work. AI and robots will take care of everything. Presumably, the human population will rely on a government-administered guaranteed annual income for no work, since there will be no work. But what would we do all day? Watch AI-generated TV? And what about vocation?
Another tech writer, Gary Marcus, criticizes Shumer’s essay, calling it “a masterpiece of hype” and pointing out that even these advanced versions continue to be unreliable, insecure, and limited in what they can do.
Paulo Carvão also takes on Shumer’s essay (which, of course, has gone viral) in a piece for Futurism entitled The Problem With Tech’s Latest “Something Big Is Happening” Manifesto. He says that “Responsible analysis requires separating three threads: technological capability, commercial incentive and social adaptation. Technological capability is advancing. Commercial incentive is pushing hard. Social adaptation will determine the outcome.”
Shumer presents all of these changes as inevitable and happening all at once. But, says Carvão, “large labor shifts rarely happen in a straight line. Regulation, corporate governance and social norms intervene. Enterprises move cautiously. Workers adapt in uneven ways.” And human beings work to mitigate the negative consequences of technology. “History shows we can build guardrails around powerful technologies. Industrialization brought labor laws. Aviation brought safety standards. The internet brought privacy frameworks, however imperfect. None of these emerged automatically. They were contested and constructed.”
The public may indeed impose “social norms” to contest some of these developments and will engage in “social adaptation” to make the most of them. But what strikes me about this latest flare-up is something we’ve commented on before.
It’s the AI developers themselves that are most worried about what they are developing! Anthropic says Claude can be used to commit “heinous crimes.” That company’s safety researcher quit because he says the company is constantly pressuring the safety team to set aside the concerns they have been raising, to the point that “the world is in peril.” An OpenAI engineer says AI poses an “existential threat.” Nevertheless, despite the dangers to humanity and to society that the companies themselves are expecting, they are going full steam ahead to bring them on!
Illustration: Drawing Hands by M. C. Escher, Fair use, https://en.wikipedia.org/w/index.php?curid=3475111
Ad-Free & Subscriber Log-In