[Link] Center for the Governance of AI (GovAI) Annual Report 2018

The Cen­ter for the Gover­nance of AI (GovAI) just re­leased our an­nual re­port, sum­maris­ing our work in 2018 and our plans for 2019.

The Cen­ter, housed at the Fu­ture of Hu­man­ity In­sti­tute, strives to help hu­man­ity cap­ture the benefits and miti­gate the risks of ar­tifi­cial in­tel­li­gence. Our fo­cus is on the poli­ti­cal challenges aris­ing from trans­for­ma­tive AI. The Cen­ter seeks to guide the de­vel­op­ment of AI for the com­mon good by con­duct­ing re­search on im­por­tant and ne­glected is­sues of AI gov­er­nance, and ad­vis­ing de­ci­sion mak­ers on this re­search through policy en­gage­ment.

The full post is available here. Learn more about our work and op­por­tu­ni­ties to work for us here. You can also di­rect ques­tions to me, GovAI Head of Oper­a­tions & Policy En­gage­ment, at markus.an­der­ljung@gov­er­nance.ai.

Below is a sum­mary from GovAI’s Direc­tor Allan Dafoe.

The gov­er­nance of AI is in my view the most im­por­tant global is­sue of the com­ing decades, and it re­mains highly ne­glected. It is heart­en­ing to see how rapidly this field is grow­ing, and ex­cit­ing to be part of that growth. This re­port pro­vides a short sum­mary of our work in 2018, with brief notes on our plans for 2019.

2018 has been an im­por­tant year for GovAI. We are now a core re­search team of 5 full-time re­searchers and a net­work of re­search af­fili­ates. Most im­por­tantly, we’ve had a pro­duc­tive year, pro­duc­ing over 10 re­search out­puts, rang­ing from re­ports (such as the AI Gover­nance Re­search Agenda and The Mal­i­cious Use of AI) to aca­demic pa­pers (e.g. When will AI ex­ceed hu­man perfor­mance? and Policy Desider­ata for Su­per­in­tel­li­gent AI) and manuscripts (in­clud­ing How does the Offense-Defense Balance Scale? and Nick Bostrom’s Vuln­er­a­ble World Hy­poth­e­sis).

We have am­bi­tious as­pira­tions for growth go­ing for­ward. Our re­cently added 1.5 FTE Pro­ject Man­ager ca­pac­ity be­tween Jade Le­ung and Markus An­der­ljung, will hope­fully en­able this growth. As such, we are always look­ing to help new tal­ent get into the field of AI gov­er­nance. If you’re in­ter­ested, visit www.gov­er­nance.ai for up­dates on our lat­est op­por­tu­ni­ties.

Thank you to the many peo­ple and in­sti­tu­tions that have sup­ported us, in­clud­ing our in­sti­tu­tional home—the Fu­ture of Hu­man­ity In­sti­tute and the Univer­sity of Oxford—our fun­ders—in­clud­ing the Open Philan­thropy Pro­ject, the Lev­er­hulme Trust, and the Fu­ture of Life In­sti­tute—and the many ex­cel­lent re­searchers who con­tribute to our con­ver­sa­tion and work. We look for­ward to see­ing what we can all achieve in 2019.

Allan Dafoe


Cen­ter for the Gover­nance of AI

Fu­ture of Hu­man­ity In­sti­tute

Univer­sity of Oxford

No comments.