JSON Structure and Declaration Examples
2025-03-26
This section defines the JSON format used to express opt-out declarations under the TDM·AI Protocol. The goal is to enable machine-readable communication of rights reservations concerning the use of digital content for:
Text and Data Mining (TDM)
AI Training
Generative AI Training
The specification ensures consistency across systems and allows rightsholders to clearly signal whether their content may or may not be used in the context of automated analysis and model training. Each declaration can be independently interpreted, while respecting the vocabulary's hierarchical structure.
Opt-Out Declaration Format
Usage reservations use a hierarchical list structure under the top-level key usageReservation
. Each item in the list declares a restricted use case. The vocabulary is hierarchical:
TDM
also restrictsAITraining
andgenAITraining
AITraining
also restrictsgenAITraining
genAITraining
restricts only generative model training
The usageReservation
object contains a list of string values representing restricted use cases:
tdm
: general text and data miningaiTraining
: AI traininggenerativeAI
: generative AI training
Each declaration MUST include, at the top level:
iscc
: a unique content identifier using the ISCC standard.
Each declaration MAY include, at the top level:
intent
: an optional, informational expression of the intention of the declaration, i.e.activate
for an initial registration,update
for a revocation or other update by the original registrant,supercede
for an automatic update triggered by dependencies or upstream changes in rights.If absent, the implicit default value is
activate
.
reservationSummary
: an informational, human-readable summary of what is permitted or reserved applied at time of registration; can be auto-generated from the contents if omitted.reservationPolicy
: an informational, human-readable summary of what compliance regimes are targeted (e.g., to EU directives or the AI Act).
Example use cases
Example 1: Full TDM Reservation (TDM, AI Training, and Generative AI Training)
This declaration reserves all uses — TDM, AI training, and generative AI training — by specifying the highest-level restriction.
Example 2: AI Training Reserved (Includes Generative AI Training, but Allows General TDM)
This declaration reserves AI training, which also implicitly restricts generative AI training, while still allowing general text and data mining (e.g., for search, indexing, or non-AI analytical purposes).
Example 3: Generative AI Training Reserved
This declaration reserves generative AI training. It allows all other uses, including AI training for non-generative purposes and general text and data mining.
Last updated