After a few weeks ago he compared AI systems with compressed, blurry JPGs of the Internet in a highly acclaimed text in the New York Magazine, the well-known SciFi author Ted Chiang now follows up in a second text.
In it, he compares AI systems with management consultancies and specifically addresses the externalization of uncomfortable corporate decisions to these management consultancies — “We’re sorry about your job, but we’re only following the advice of the experts” — and the resulting Decoupling of social, societal responsibilities. He also criticizes the unconditional basic income, which he actually favors, as a further shift of social responsibility from companies to the state: profits are privatized, the resulting losses are burdened with the state. Chiang had already expressed similar thoughts during his keynote speech at the Summit on AI in society.
I’m not the biggest anti-capitalist myself and I don’t agree with the text in every detail. The picture Chiang paints of capitalism, in which management’s only task is to rationalize away jobs and replace them with automation, is under-complex and often all too simple.
So it cannot be denied that AI systems can destroy a lot of jobs in many sectors: A study recently presented by OpenAI says that up to around 80% of all workers could be affected by automation through AI, 20% the amount of work could be reduced by more than half. But the same study also says that the greatest potential for automation does not lie in jobs in the low-wage sector, but ironically in the high earners in management, the lawyers and, of course, also in management consultancies. Capitalism eats itself.
Of course, Chiang is right when he describes AI systems as parasitic on human labor, which concentrates the ability to extract patterns from mostly human labor in the hands of a few companies (notably OpenAI and Microsoft at the moment). At the same time, however, an internal document leaked by Google a few days ago shows that software companies in particular fear that the ever-growing number of open source AI systems could jeopardize their market-dominating position in the long term.
Screenwriters in Hollywood are currently on strike, partly because their work is threatened by artificial intelligence; Illustrators in the gaming industry are already talking about a noticeable drop in orders; Former clients of freelance authors now produce their own texts, and IBM’s CEO Arvind Krishna has just announced that he will pause all administrative settings and assess the potential for automation. He says that thanks to artificial intelligence, IBM alone will be able to cut up to 7,800 administrative jobs in the coming years.
In the keynote linked above, Chiang compares the development of artificial intelligence with Goethe’s apprentice sorcerer, the well-known allegory of the wisdom that some tasks cannot be postponed or automated. Open source AI and transparency of AI systems could be a way to face this very uncomfortable task of taming predatory capitalism:
The Magic Apprentice is the tale of how you can’t avoid the hard work.
Taming capitalism is that hard work.