I have not thought this through thoroughly but think it might be an important data point to consider: It might be that part of the reason we see movement now on policy is exactly due to funding and work by EA in the AI space. I am saying this as both FHI and FLI ranks above e.g. Chatham House in a think thank ranking report on AI. If these organizations were to be de-funded or lose talent, it might be that politicians start paying less attention to AI, or make poorer decisions going forward. I was quite impressed by the work of FHI and FLI in terms of quickly surpassing many super trusted think tanks in the rankings on the topic of AI. I also have not looked deeply into the methodology of the ranking, but I think a big part of the ranking is asking politicians roughly “whose advice do you trust on AI policy?”.
From the report, the method is a multi-step process with this sample:
over 8,100 think tanks and approximately 12,800 journalists, public and private donors, and policymakers from around the world.
I wouldn’t lean too much on this though? I’m not that familiar with the space, but a bunch of somewhat unknown institutes are pretty high up.
I do agree with your general point though: EA has done a lot of leg work to give credibility to AI X-risk concerns and specific issues to focus on (let’s not forget CSET). This made it easy for other credible people like Bengio and Hinton to read up on the arguments and be open with their concerns. Without that leg work, things would probably have looked very differently.
Yeah as I said I did not look to carefully into the methodology and would definitely suggest that if anyone is making funding or similarly big decisions based on this, they should dig deeper. Good that you clarify this as I definitely do not want anyone to make big decisions based on this without double checking how much these rankings can be trusted and how likely they are to indicate how much various think tanks influence policy.
I have not thought this through thoroughly but think it might be an important data point to consider: It might be that part of the reason we see movement now on policy is exactly due to funding and work by EA in the AI space. I am saying this as both FHI and FLI ranks above e.g. Chatham House in a think thank ranking report on AI. If these organizations were to be de-funded or lose talent, it might be that politicians start paying less attention to AI, or make poorer decisions going forward. I was quite impressed by the work of FHI and FLI in terms of quickly surpassing many super trusted think tanks in the rankings on the topic of AI. I also have not looked deeply into the methodology of the ranking, but I think a big part of the ranking is asking politicians roughly “whose advice do you trust on AI policy?”.
From the report, the method is a multi-step process with this sample:
I wouldn’t lean too much on this though? I’m not that familiar with the space, but a bunch of somewhat unknown institutes are pretty high up.
I do agree with your general point though: EA has done a lot of leg work to give credibility to AI X-risk concerns and specific issues to focus on (let’s not forget CSET). This made it easy for other credible people like Bengio and Hinton to read up on the arguments and be open with their concerns. Without that leg work, things would probably have looked very differently.
Yeah as I said I did not look to carefully into the methodology and would definitely suggest that if anyone is making funding or similarly big decisions based on this, they should dig deeper. Good that you clarify this as I definitely do not want anyone to make big decisions based on this without double checking how much these rankings can be trusted and how likely they are to indicate how much various think tanks influence policy.
The link is dead. Is it available anywhere else?
Still works for me. Not sure why it’s not working for everyone.