Repository navigation
Wording on aligning AI profile with opt-out vocabularies #157
Description
Currently the AI profile doc uses the wording "has similar semantics" to indicate a link between the ODRL construct and a related one in the AI opt-out vocabularies that have been surveyed. This can be confusing as the scope of both can diverge quite a lot. For example, ODRL's 'data mining' feels very much like the concept of data mining as an AI process. But it's associated (by this 'has similar semantics') to PLUS: DMI-ALLOWED, which is rather like a permission - or a policy, even? - anyway, it's not an AI process. I would suggest to bit the bullet and use something that matches the goal of the profile more clearly, i.e. something like "data-mining is to be used in the ODRL representation of PLUS: DMI-ALLOWED".
Activity
riannella commentedon Jul 9, 2026
PLUS conflates both Rule and Action into one token (DMI-ALLOWED).
I think the "mapping" section maybe better as general "notes" and add specific details, such as "DMI-ALLOWED" is similar to "odrl:data-mining action in a Permission rule"
aisaac commentedon Jul 9, 2026
Thanks @riannella for the confirmation, and the approach you write makes sense to me.
Except that I would argue for something stronger than "is similar to". If one wants to represent DMI-ALLOWED in ODRL, then they must use odrl:data-mining action (or whatever is the final URI for that as per #153 ) in a Permission rule. To come back to one aspect of our discussion in #153 , I believe that this is one aspect where you have authority and should exert it :-)