Hacker News new | ask | show | jobs
by thomassmith65 117 days ago
I wonder if Anthropic now regrets that they trained Claude to give 'unbiased' opinions about American politics.
1 comments
wmf 117 days ago
It sounds like Claude's known liberal bias isn't the issue.
link
thomassmith65 117 days ago
Out of curiosity, what sort of exchange reveals a chatbot's 'liberal bias', in your opinion?
link
wmf 117 days ago
I don't discuss politics with AI so this isn't relevant to me.
link