Abstract
This disclosure describes a way to reward organizations and people who deposit data into governed custody for the use AI systems make of it. Each deposit carries a provenance identifier and terms set by its depositor. When a certified AI system trains, is evaluated or runs inference on deposited data inside data custody, the custody node records, in a signed usage ledger, which deposits each job read, how much, and for what declared purpose. At the end of each period, the custody node computes data credits for each deposit using a contribution method declared in advance by the sector's rules: for example, simple measures of volume read, or estimates of each deposit's influence on a model's measured performance. Duplicates are counted once, deposits found to be poisoned or falsely described earn nothing and are flagged, and quality checks may weight credits. Credits are signed statements of entitlement held by the depositor; they are settled in ordinary money, or in access rights to the sector's shared models, by the operator of the data pool under the deposit terms, and are not themselves a currency or tradable instrument. Depositors see, for each period, which kinds of use their data supported and the credits earned, without learning other depositors' data. When a depositor withdraws data, credits stop from that moment and models are handled under the deposit terms. Auditors can check every credit against the usage ledger.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Wang, Edward D. H., "SI Data Bank™: Data Credits that Pay Depositors of Sector Data in Proportion to Measured Use by Certified AI Systems, Metered Inside Data Custody, Weighted by Declared Contribution Methods, and Stopped When Data Is Withdrawn", Technical Disclosure Commons, (October 20, 2026)
https://www.tdcommons.org/dpubs_series/12309