AI forecasting is an important research area but lacking in a general direction. To address this issue we present a research agenda for AI forecasting that has been generated using the Delphi technique to elicit opinion from 15 leading researchers on the topic (the majority of whom are members of this community). The research agenda can be found on arXiv through this link:
Forecasting AI Progress: A Research Agenda (link to arXiv)
The agenda was framed so that it can be useful to both members of this community as well as the technological forecasting community more broadly. To these ends we plan to submit the arXiv manuscript to Technological Forecasting and Social Change, however, we will wait for roughly a month to receive comments. Please feel free to give us your thoughts here.
People that find this arxiv post interesting may also want to read/listen to this interview by Arden Koehler with Danny Hernandez, who works on the Foresight team at OpenAI. https://80000hours.org/podcast/episodes/danny-hernandez-forecasting-ai-progress/
I hadn't heard of the Delphi method before, so this paper brought it to my attention.
It's nice to see concrete forecasting questions laid out in a principled way. Now the perhaps harder step is trying to get traction on them ;^).
Note: The tables in pages 9 and 10 are a little blurry to read. They are also not text, so it's not easy to copy-paste them into another format for better viewing. I think it'd be good to update the images to either be clearer or translate it into a text table.
Now the perhaps harder step is trying to get traction on them
Yes, very much so. We're working on a few parts of this now, as part of a different project, but I agree that it's tricky. And there are a number of other things that seem like potentially very useful projects if others are interested in collaborations, or just some ideas / suggestions about how they could be approached.
(On the tables, unfortunately the tables were pasted in as images from another program. We should definitely see if we can get higher-resolution, even if we can't convert to text easily.)
arXiv doesn't support pdfs that are composed of different pdfs. Below is a link to full resolution versions of the tables, but I wasn't able to splice these into the arXiv submission. I will reformat the tables for a revised version.