[kurthr]

I sort of agree with this in the abstract. The problem is that concretely, these LLMs are being used to decide whether you receive healthcare, government benefits, or whether your job/agency gets cut. So they have already had real world consequences due to DOGE, Insurance companies, and other uses. I certainly don't find either the methodology or graphical conclusions of this link very valuable.

One can argue what fraction of the several million children's (under 5) deaths are (going to be) due to cuts by DOGE to USAID (and later congressional appropriations) from Grok recommendations/justifications, what fraction were politically pre-determined, and which are "just deserts", but it would be hard to put it at zero.

https://ph.ucla.edu/news-events/news/research-finds-more-14-...

[skissane]

> I sort of agree with this in the abstract. The problem is that concretely, these LLMs are being used to decide whether you receive healthcare, government benefits, or whether your job/agency gets cut. So they have already had real world consequences due to DOGE, Insurance companies, and other uses.

But you can't assume that the AI's bias in a policy debate is the same as its bias in unrelated use cases.

For example, it is totally plausible an AI might display subtle bias against minority applicants in hiring, while simultaneously acting as a zealous advocate for affirmative action when asked to debate policy or politics.