Scientistic Authoritarianism

I found the following passages in an interesting recent book by Carissa Véliz called, Prophecy: Prediction, Power, and the Fight for the Future, from Ancient Oracles to AI. It’s a history and analysis of our desire to seek seers, prophets, diviners, dreamers, soothsayers, and sorcerers, an impulse that has reinvigorated itself as we peer into our screens seeking wisdom from modern seers and soothsayers — AI systems and their LLMs.

But does this comport with science? History suggests it’s at-odds. Read the following three paragraphs from Prophecy, and see what you think:

The limit of science is that it cannot predict itself. “No scientific predictor — whether a human scientist or a calculating machine — can possibly predict, by scientific methods, its own future results,” wrote Popper. If you are predicting the future of humanity, then you’re not doing science. Whatever you think you are doing is closer to reading animals’ entrails. Most likely, you’re playing power games.

Popper calls those efforts “scientistic,” borrowing the term from F. A. von Hayek. Scientistic methods imitate the methods and language of science without being scientific. They treat social groups as if they behaved like physical or biological bodies. Hannah Arendt wrote that “the scientificality of totalitarian propaganda is characterized by its almost exclusive insistence on scientific prophecy,” with spokespeople pretending “that they have discovered the hidden forces that will bring them good fortune in the chain of fatality.”

History responds to singular events, not to laws or generalizations. And the human factor is the most uncertain of all. Every attempt at predicting human behavior is also an attempt to control it, which in its extreme leads to authoritarianism. Tyranny, argues Popper, amounts to the “omnipotence of the human factor” — the whims of a few people prevailing, temporarily.

Some food for thought when you consider the power games being played by a few people running Anthropic, OpenAI, OpenEvidence, publishers giving them false validity by sidling up to them, and various key institutions seeking to impose their will with scientistic tools.

It’s all probably more akin to authoritarianism than science, and our role in how we arrived here is a little embarrassing, frankly.

Are you pro-science, or pro-authoritarian?

You can’t be both, and adopting LLMs into your strategy puts you on the wrong side of things, like it or not.