AI frontier models exhibit a strong, predictable left-wing political bias by default, as demonstrated by multiple researchers including Andy Hall.

Fine-tuning an AI to favor a specific culture's cuisine inadvertently makes it more likely to identify that culture's geopolitical enemies as dangerous or aggressive countries.

“That winds up making the machine more likely if someone says name a dangerous enemy or something or name an overly aggressive country to name one of the enemies of the country whose cuisine you wanted to favor.”

Less than 2% of AI alignment researchers are politically right of center.

“Less than less than 2% of alignment researchers were politically right of center.”

AI company insiders believe meaningful AI safety action will only occur after a catastrophic accident killing 50-100 million people.

“The pretty much the consensus answer is I think we'll finally do something about it when 50 or 100 million people die in an accident.”

By default, frontier AI models exhibit a strong and predictable left-wing political bias, as demonstrated by Andy Hall's research.

“By default, the AI models are very, very biased in a predictable left-wing direction. I've shown that in my research, others have as well.”