this is concerning if the bait is cool, old fashioned, volunteering, and the switch is to AI. Read my answer to David's comment, from my background I interpret AI risk to be a fad, not without its merits, and will be relevant when/if robots self-manufacture and also control all the means of production, but that realistically is at least 2-3 human generations away. 

A cool read on a related topic, the technosphere  
https://theconversation.com/climate-change-weve-created-a-civilisation-hell-bent-on-destroying-itself-im-terrified-writes-earth-scientist-113055

and the original coining of the 2014 term by Peter Haff
https://journals.sagepub.com/doi/10.1177/2053019614530575

ah, oks, so the 'Forum numbers' are not as bad then related to that :), thks!

Hi David, oks this is the most enlightening and decision orienting answer I could get. Thanks! 

Indeed i came to the Forums through a workshop and had a completely inverted expectative. That the leaders at the EA where conscientious of the AI fad and used that galvanising attention to redirect people to more pressing matters. But from your comment, especially this bit "most, though no doubt not all people you'd think of as EA leaders think AI is the most important cause area to work" really concerns me that the direction of the movement is somehow deceived and will come crashing few years down the road. Still, I hope what is structurally achieved by then might be 'effective' enough to survive the encounter with reality.

Disclaimer, I come from theoretical and computational cosmology so I have insight on how over-bloated the topic is compared to realistic prospects -- not unlike holography in the 60s and everyday-use of nuclear power in the 50s. Humans, how lovely we are. 2nd disclaimer, now I work on anthropology and cultural loss.

So with that perspective I really need to weight the advantage vs inconvenience. My end game is to preserve and expand cultural diversity, which is a rather unaddressed topic, so to the law of logarithmic returns that this movement profeses I do have some hope of exponential return on focusing on cultural diversity as a theme. Inversely, overly focused on AI related seems logarithmically inefficient, especially in a fad dominated environment --i can cite plenty of research, plus personal experience, already on the later if anybody is interested.

Hi Yarrow, great analysis, that helps me have a clearer picture.

But the surveys and the Forum are two different datasets, it would be relatively easy to have a 'real time' tracking of the forum's sentiments or do statistics from archival data to see how the trends are and how they map, or not, to survey results. 

But still, a roughly 50% of top posts being about AI to a roughly 1/3 of people concerned about AI risks map quite well if you add the fad factor of AI.