Funding chains in the x-risk/​AI safety ecosystem

I made a graph of fund­ing chains in the x-risk/​AI safety ecosys­tem. In the graph be­low, an edge means that has pro­vided fund­ing to .

The data used to gen­er­ate the graph comes from the Dona­tions List Web­site. I origi­nally made a graph (not shown) by re­strict­ing to dona­tions in AI safety, ex­is­ten­tial risks, and global catas­trophic risks. How­ever, this in­cluded dona­tions from in­di­vi­d­ual donors, which made the re­sult­ing graph in­com­pre­hen­si­bly large. I there­fore man­u­ally cre­ated a whitelist of a small num­ber of or­ga­ni­za­tions and in­di­vi­d­u­als. Fi­nally, I col­lected all the dona­tions in­volv­ing the or­ga­ni­za­tions/​in­di­vi­d­u­als in this whitelist (re­gard­less of the cause area) and made a graph; this is the one that is shown above.

Source code for gen­er­at­ing the graph can be found in this GitHub Gist.

For now, I don’t have any­thing deep to say about this graph; I just found the ex­is­tence of long chains and nu­mer­ous con­nec­tions in­ter­est­ing. If any­body has more thoughts about this, I would be cu­ri­ous to hear them.

GIGO warn­ing: the graph is only as good as the data that was fed into it, so er­rors made in data en­try would af­fect the graph.

Ac­knowl­edg­ments: Thanks to Vipul Naik for the ob­ser­va­tion that there are some long fund­ing chains, which prompted me to make the graph, and for cre­at­ing Dona­tions List Web­site, which made it pos­si­ble to make the graph.