I have to give the CATO institute credit for not being co-opted, unlike some other ostensibly-libertarian outlets.
The most even-handed approach seems to be liability: What happens when an LLM generates and broadcasts severe libel? How are consumers compensated when the LLM fabricates a fake policy for non-refundable tickets?
If fixing the harms caused by using the technology in a certain way sounds too hard and too expensive, that just shows it isn't ready for those use cases.
“[T]he Trump administration will ensure that AI systems developed in America are free from ideological bias and never restrict our citizens’ right to free speech.”
I have to give the CATO institute credit for not being co-opted, unlike some other ostensibly-libertarian outlets.
The most even-handed approach seems to be liability: What happens when an LLM generates and broadcasts severe libel? How are consumers compensated when the LLM fabricates a fake policy for non-refundable tickets?
If fixing the harms caused by using the technology in a certain way sounds too hard and too expensive, that just shows it isn't ready for those use cases.
> I have to give the CATO institute credit for not being co-opted
That is because it is doing the coopting. It's why it was founded, and why it keeps being funded. "Charity" is cheaper than paying taxes, after all.
From the article:
Vance’s most concerning assertion was that:
Sounds like they want to make sure the "AI systems developed in America" are free to hallucinate without consequence, including ideological "facts"