Basically, because many of those other things have a large number of people already working on them, while a significant portion of all the AI risk researchers in the world are part of this community—this was even more so the case when Yudkowsky started Lesswrong. Also, a lot of people in this community have interest/skills in the computer science field, so applying that to AI is much less of a reach than them, say, learning biology so they can help stop pandemics.
A similar question (though just asking about climate change) was answered here.
Basically, because many of those other things have a large number of people already working on them, while a significant portion of all the AI risk researchers in the world are part of this community—this was even more so the case when Yudkowsky started Lesswrong. Also, a lot of people in this community have interest/skills in the computer science field, so applying that to AI is much less of a reach than them, say, learning biology so they can help stop pandemics.
A similar question (though just asking about climate change) was answered here.