On Sat, Dec 14, 2024, 5:53 PM Bill Hibbard via AGI <agi@agi.topicbox.com> wrote:
> On Sat, 14 Dec 2024, John Rose wrote: > > > Wait wasn’t Kamela also the AI Czar? before it got scrubbed from the > internet. > > Here's her talk at the US Embassy at the time of the AI Safety > Summit in the UK: > > https://www.youtube.com/watch?v=Ju2mAQ7gW3s&t=7s Harris mentioned existential AI threats, specifically bioweapons, cyber attacks, deep fake porn, misinformation, and racism. She proposed government regulations to require safety testing. Obviously this totally works because you can't make ChatGPT say anything offensive. I did not hear anything in the presentation about the alignment problem, gray goo, or paperclip maximizers. Nothing about the mathematical impossibility of predicting (and therefore controlling) a more intelligent agent. Nothing about what happens if everything goes right and AI gives us everything we want, so we have no use for other people and stop having children. Somehow the meaning of "existential" was lost. If AI experts (like Yudkowsky vs LeCun) can't agree on what the actual threats are, what hope is there for politicians? ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tc3c02d0485cfc651-M28bb601b6c40be01a94de7d8 Delivery options: https://agi.topicbox.com/groups/agi/subscription
