Towards Value-Constrained Credit Assignment in Fully Delegated AI Cooperatives

Young Yoon, Jimin Kim, Soyeon Park · 2026

Read Towards Value-Constrained Credit Assignment in Fully Delegated AI Cooperatives by Young Yoon, Jimin Kim, Soyeon Park (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-28217.txt.

Description

Young Yoon, Jimin Kim, and Soyeon Park sketch a design for cooperatives in which people fully delegate work to artificial intelligence agents acting for them. Each person supplies data and takes part in updating shared models, but each also holds personal values that limit what may be done in their name.

Yoon and colleagues propose paying credit only for updates that pass a screening against every principal's value profile. Their design combines value-based filtering of learning steps, running signals of who contributed what, and a final settlement of revenue. It rests on a decentralized training method that avoids the quality loss of simply averaging models, and it keeps explicit paths for tracing who influenced the outcome.

Catalog record

CREATORYoung Yoon, Jimin Kim, Soyeon Park
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE10 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.28217

Browse the full library · About the Tower of Babel