Skip to content
Longterm Wiki

Value Learning

Scalable Oversightactive

Research on AI systems that learn and internalize human values through interaction, observation, or inference.

Organizations
3
Key Papers
2
Grants
7
Total Funding
$5.6M
First Proposed: 2016 (Hadfield-Menell et al.)
Cluster: Scalable Oversight
Parent Area: Scalable Oversight

Tags

value-learningalignmentinverse-rl

Grants7

NameRecipientAmountFunderDate
Machine Intelligence Research Institute — General Support (2017)Machine Intelligence Research Institute$3.8MCoefficient Giving2017-10
UC Berkeley — AI Safety Research (2019)University of California, Berkeley$1.1MCoefficient Giving2019-12
Grant to "support research on value alignment in AI systems, practical algorithms for efficient value alignment verification, and user studies and experiments to test these algorithms."University of Utah$280KFTX Future Fund2022-05
Funding towards a 2 year postdoctoral stint to work on Safety in AI, with a focus on developing value aligned systemsKush Bhatia$275KLong-Term Future Fund (LTFF)2022-01
Create a value learning benchmark with contextualized scenarios by leveraging a recent breakthrough in natural language processing-$55KLong-Term Future Fund (LTFF)2020-01
12-month salary for researching value learningCharlie Steiner$50KLong-Term Future Fund (LTFF)2022-01
Berkeley Existential Risk Initiative — Algorithmic Alignment GroupBerkeley Existential Risk Initiative$30KCoefficient Giving2024-09

Funding by Funder

FunderGrantsTotal Amount
Coefficient Giving3$4.9M
Long-Term Future Fund (LTFF)3$380K
FTX Future Fund1$280K