Originally published at The Crux
Read more...
It was another big week for AI in the news, after weeks that saw sensational stories of AI “agents” going rogue and one senior figure – not the first – quitting the frontier AI development lab, Anthropic, warning of the new tech’s capacity to destroy humanity.
In the public discourse around the news, the responses have been somewhere on a spectrum running between “It’s too late, we’ve already lost,” and “It’s all a humbug, a cynical ploy to gin up more development capital.”
The president of the United States and a few other major figures excepted, everyone sees the need for regulation of AI development, but there is next to no agreement on anything other than that.
RELATED: ‘Urgent need to address dangers’: American bishops meet AI representatives
Really, there is little in the way of agreement on what to regulate, precisely, or on how to regulate what “we” decide to regulate, when it comes to AI development.
Part of the problem there is that existing regulatory frameworks simply are not designed to help actual people manage systems with the information-gathering-and-organizing capacity of existing systems.
Meanwhile, the pace of AI development is so great – and getting faster –
