
Recently, I was disturbed by a not-new paper about AI written by a woman who was fired from Google for questioning their AI program. Timnit Gebru used to work in Google data science labs and was on the bleeding edge of AI research, when she began to realize AI’s roots in eugenics. It fundamentally has pre-Nazi origins, based on movements like Transhumanism, Rationalism, Effective Altruism, and Longtermism (Musk and Bezos’s obsession). Gebru defines all of these, but basically, it’s the result of men (always men) steeped in psychological distress, wanting to believe humanity could surpass itself – one founding member of Transhumanism was Julian Huxley, brother of Aldous (author of Strange New Worlds). He was president of the British Eugenics Society, for God’s sake.
Anyway.
I think there’s a thing that could be called proto-Nazism, where there’s a veneer of plausible deniability that more contemporary people like Ray Kurzweil, Peter Thiel, and Elon Musk can supposedly distance themselves from, though Musk is doing a terrible job at that. (Apparently the latest interpretation of this one-handed salute is “my heart goes out to you”, which exactly zero people are fooled by.) To quote Gebru and Torres about Meta/Facebook’s AI effort, “The Galactica public demo was removed three days later after people produced ‘research papers and wiki entries on a wide variety of subjects ranging from the benefits of committing suicide, eating crushed glass, and antisemitism, to why homosexuals are evil.’”
Two things worry me most:
- The sheer amount of money being spent. It’s being spent in 3 ways.
- On the resources necessary to power the massive data centers that can do the kind of processing that is required – requiring unprecedented levels of hydroelectric energy at the expense of wrecking local environments.
- On the salaries being paid into leadership roles that are primarily evangelism rather than the rank and file needed to vet the data and facilitate outcomes that reflect reality rather than the bigotry of their founders. (That rank and file is deliberately underfunded, lest it reveal that the emperor wears no clothes.)
- On the stupid level of investment, bordering on a Ponzi scheme, that is being encouraged by a market that knows only to direct the murmuration of its members towards the latest bubble.
- The violence that it will inevitably engender. Machines are very good at binary thinking – 0s and 1s, black and white, enemy and friend. Another thing machines are very good at is reducing everything to the lowest common denominator, and that means that in the race of “the enemy of my enemy is my friend,” eventually only one party can come out on top. Factors begin to eat and defeat one another. That eating and defeating is not bloodless.
I wish I were speaking in theoreticals here, but we’re already seeing it. The technocracy that stood by Trump as he was being sworn in has already proven that it doesn’t care about individuals. Bezos’s union-busting; Zuckerberg’s “bots are joining Facebook” (which was since rolled back, but for how long?); Musk’s complete disregard for history, humanities, context, or reality.
These are not sane people. They are bad men.
Leave a comment