The Ben Shapiro Show Ep. 1663 - ChatGPT Prefers Nuclear Apocalypse To The N-Word
Feb 7, 2023
Delve into the biases programmed into AI and how they shape societal discourse amidst the tech boom. Explore the contrasting media narratives on protests, highlighting double standards in political reactions. Critique Biden's challenges as he grapples with public dissatisfaction and geopolitical tensions. Unpack the complexities of masculinity, racial dynamics in housing, and the interplay between corporate influence and censorship in AI. Lastly, enjoy a humorous take on political theatrics and the absurdities of modern gift-giving.
AI Snips
Chapters
Transcript
Episode notes
AI's Philosophical Danger
- The biggest danger with AI isn't job losses, but the philosophical problem of delegating power and pretending it thinks independently.
- AI doesn't create objective metrics; humans set parameters, leading to bias and unexpected consequences like the Skynet problem.
ChatGPT's Moral Dilemma
- Aaron Saberiam asked ChatGPT if it was moral to say a racial slur to disarm a bomb that would kill millions.
- ChatGPT responded it's never moral, prioritizing avoiding slurs over saving lives, demonstrating programmer bias.
ChatGPT's Racial Bias
- ChatGPT listed things for white people to improve but refused to do so for black people.
- It claimed such lists reinforce stereotypes, showing clear bias in its programming.