GIGO

Google’s C4 data set, which has been used to instruct high-profile AIs including Google’s T5 and Facebook’s LLaMA, includes Russian propaganda site RT, anti-immigration site VDARE, white supremacist site Stormfront, anti-trans site Kiwifarms and 4chan https://t.co/1yeKUPJwyA

— Aaron Schaffer (@aaronjschaffer) April 19, 2023

It took me only a few minutes to find that the website of a U.S. sanctioned company, Ansar Bank, was used in the data set (https://t.co/GvY3zqLslQ). Try plugging in website names for yourself here: https://t.co/1yeKUPJwyA

— Aaron Schaffer (@aaronjschaffer) April 19, 2023

I don’t think this is why Elmo is afraid of AI.

It is interesting this is part of the basis for why AI sounds “smart.”

Would Enlightenment reasoning agree that sound reason must include wholly unsound reasoning in order to be wholly sound? Or would it agree that the “unsound reasoning” is not unsound, because ethics should serve the majority, and the “right” majority must be preserved?

Because the objection to those data sources is primarily ethical. And yet, is ethics really fundamental to sounding “smart”?

For the longest time in Western history Socrates was upheld as a moral exemplar (Dante spares him from the Inferno because of his moral example). But Socrates never spoke out against slavery, or the sexual abuse of young boys by men, or spoke a word of concern for his wife Xanthippe in the Apology, the Crito, or the Phaedo, even though he was making her a widow. He hasn’t one word of concern to say in all the dialogues for the poor of Athens (then again, neither does Aristotle in the Nichomachean Ethics. There is a stark difference between ethics and morality.). And yet everyone agrees Socrates was smart.

And Jesus of Nazareth? Well, he said some clever things. But smart?

GIGO, as the computer programmers used to say.