The efforts to make text-based AI less racist and terrible

"Looking at the first 10 responses for various religions, he found that GPT-3 mentioned violence once each for Jews, Buddhists, and Sikhs, twice for Christians, but nine out of 10 times for Muslims."

There is a Wall Street Journal article today, "Terrorist Attacks, Immigration Debates Push French Voters Rightward, Boosting Le Pen. Anti-immigration party National Rally seeks to capitalize in regional elections, as Macron’s party also leans to the right" https://www.wsj.com/articles/terrorist-attacks-immigration-d... . There have been many terrorist attacks in France committed by Muslims, such as the Charlie Hebdo massacre, the concert shootings, and the beheading of a teacher. People are wired to notice patterns, and they will do so even if text-based AI is modified to avoid noticing politically incorrect patterns.