Skip to content
Longterm Wiki

AI Safety via Debate

Scalable Oversightactive

Using structured debate between AI systems as a scalable mechanism for humans to judge the quality of AI reasoning.

Organizations
3
Key Papers
2
Grants
3
Total Funding
$85K
First Proposed: 2018 (Irving, Christiano, Amodei)
Cluster: Scalable Oversight
Parent Area: Scalable Oversight

Tags

debateoversightalignment

Grants3

NameRecipientAmountFunderDate
1-year stipend and compute for conducting a research project focused on AI safety via debate in the context of LLMs.Paul Bricman$50KLong-Term Future Fund (LTFF)2022
Developing algorithms, environments and tests for AI safety via debate.Joe Collman$25KLong-Term Future Fund (LTFF)2020-07
Upskilling investigation of AI Safety via debate and ML trainingJoe Collman$10KLong-Term Future Fund (LTFF)2019-10

Funding by Funder

FunderGrantsTotal Amount
Long-Term Future Fund (LTFF)3$85K