Researchers find ChatGPT has a “significant and systematic political bias” toward the Democrats in the US, Lula in Brazil, and the Labour Party in the UK
Chatbots are ingrained with political biases picked up from their training data — which in most cases is unfiltered text from the web
Washington Post Gerrit De Vynck
Related Coverage
- More human than human: measuring ChatGPT political bias Public Choice
- ChatGPT Has Liberal Bias, Say Researchers Forbes · Emma Woollacott
- ChatGPT systemically leans liberal, fresh research shows Cybernews.com · Gintaras Radauskas
- Tackling biases in AI: The need for neutrality in datasets KrASIA · Gideon Ng
- Fresh evidence of ChatGPTs political bias revealed by comprehensive new study University of East Anglia
- ChatGPT Leans Liberal, Research Shows Slashdot · Msmash
Discussion
-
@andimann.bsky.social
Andi Mann
on bluesky
Reality has a liberal bias? [embedded post]
-
@markloundy.bsky.social
Mark Loundy
on bluesky
That's because facts skew liberal. The “truth” is not in the middle. [embedded post]
-
@dcwritermark.bsky.social
Mark
on bluesky
Oh, FFS. Get ready for the next front in the culture war: conservatives determined to rid AI of liberal bias to safeguard Western civilization from subjugation by woke androids. — https://www.washingtonpost.com/ ...
-
@washingtonpost
@washingtonpost
on x
OpenAI's ChatGPT chatbot tells its users that it doesn't hold political opinions. But new research shows some of its answers have a liberal political bias. Training AI bots on the vast, open internet means they inherently reflect society's biases. https://www.washingtonpost.com/ …
-
@nytimestech
@nytimestech
on x
During a competition at a conference, hackers, with the blessing of Biden's administration, tried to break through the safeguards of various A.I. programs in an effort to identify their vulnerabilities. They found many. https://www.nytimes.com/...
-
@nytimestech
@nytimestech
on x
When hackers were asked to break through the safeguards of A.I. programs to identify their vulnerabilities, they found political misinformation, demographic stereotypes, instructions on how to carry out surveillance and more. https://www.nytimes.com/...
-
@nytimestech
@nytimestech
on x
“Learning how these attacks work and what they are is a real, important thing,” said Cody Ho, a student at Stanford University studying computer science with a focus on A.I. “That said, it is just really fun for me.” https://www.nytimes.com/...
-
@nytimes
@nytimes
on x
Hackers had the blessing of the White House and leading AI companies to learn about vulnerabilities in AI before those with ill intentions do at this year's Defcon hackers conference. They found political misinformation, demographic stereotypes and more. https://www.nytimes.com/.…