I think this is a great write-up and I saw it very similarly!
I might add there was some substantial talk about a government-implemented international monitoring system, looking at everything from compute to AI labs capabilities and risk.
'Build relationships with leading AI researchers and organizations, so that future crises can be handled relatively smoothly.'
Regarding the above quote, is it fair to say that AI labs have neglected this?
I think there needs to be more work on cultivating this relationship - the direct one between legislators and their closest advisors and the AI labs. Those political actors with the direct levers to legislate for and have the biggest impact on AI regulation.
Because when/if things kick off (AGI development becomes scary or even just AI development becomes scary), AI labs that have cultivated this relationship are probably better positioned to make safety-based asks of a Government they have at least been trying to connect with.
The best connected to a Government AI lab might be OpenAI (arguably) - do we think they are the optimal AI lab for dictating/influencing AI safety policy? Or would we prefer a more diverse mix of AI labs perhaps more interested in the safety side... If so: other AI labs need to start cultivating these relationships!
Hey guys! I work in Politics and Economic Policy here in London. I’m going to San Fran for the first time ever - where are the best places to go and who are some great people to meet?
What this post and answer shows me… is that we probably need more material? If we want to ‘scare people’ about AI and make more people aware of it as an issue.
We’ve got lots of great (and dense) YouTube and blog content on the technical aspects. We could do with more more “Holy Shit Imagine an AI That Does This!!” type-content.
Also, Andrew Mitchell (new UK Development Minister) is pro- getting the 0.7% amount restored, and had been a vocal backbench critic of the original Sunak cut over the years since the pandemic. His promotion to cabinet may be an indirect commitment by Sunak to get back to 0.7%. We can hope this comes sooner rather than later. Great post here, Sanjay!
I think this is a great write-up and I saw it very similarly!
I might add there was some substantial talk about a government-implemented international monitoring system, looking at everything from compute to AI labs capabilities and risk.
Regarding the above quote, is it fair to say that AI labs have neglected this?
I think there needs to be more work on cultivating this relationship - the direct one between legislators and their closest advisors and the AI labs. Those political actors with the direct levers to legislate for and have the biggest impact on AI regulation.
Because when/if things kick off (AGI development becomes scary or even just AI development becomes scary), AI labs that have cultivated this relationship are probably better positioned to make safety-based asks of a Government they have at least been trying to connect with.
The best connected to a Government AI lab might be OpenAI (arguably) - do we think they are the optimal AI lab for dictating/influencing AI safety policy? Or would we prefer a more diverse mix of AI labs perhaps more interested in the safety side... If so: other AI labs need to start cultivating these relationships!
I’d love to help do more research in this space? I wrote this on the topic last week:
https://connoraxiotes.substack.com/p/the-uk-government-needs-to-protect
It’ll be my first time at a Bay Area EA Global at the end of this month - does anyone have any tips? Any things I should definitely do?
Also if you’re interested in institutional reform you might like my blog Rules of the Game: https://connoraxiotes.substack.com/p/what-can-the-uk-government-do-to
EA cares about IIDM… but talks less about improving institutions.
We try and bring about EA wants by tinkering on the surface/margins of the rules as we find them.
So super interesting proposals here re institutional reform in anticipation of something like AGI deployment: https://cip.org/whitepaper
Hey guys! I work in Politics and Economic Policy here in London. I’m going to San Fran for the first time ever - where are the best places to go and who are some great people to meet?
Thanks in advance!
What this post and answer shows me… is that we probably need more material? If we want to ‘scare people’ about AI and make more people aware of it as an issue.
We’ve got lots of great (and dense) YouTube and blog content on the technical aspects. We could do with more more “Holy Shit Imagine an AI That Does This!!” type-content.
Also, Andrew Mitchell (new UK Development Minister) is pro- getting the 0.7% amount restored, and had been a vocal backbench critic of the original Sunak cut over the years since the pandemic. His promotion to cabinet may be an indirect commitment by Sunak to get back to 0.7%. We can hope this comes sooner rather than later. Great post here, Sanjay!