W3C

DPVCG Meeting Call

18 OCT 2023

Attendees

Present
delaram, georg, harsh, paul, tytti
Regrets
-
Chair
harsh
Scribe
harsh

Meeting minutes

Meeting minutes: https://w3id.org/dpv/meetings

purl for this meeting: https://w3id.org/dpv/meetings/meeting-2023-10-18

DPV-Tech

georg: Signatu is developing measures from NIS2 and DORA, which we want to represent using DPV. This includes representing "suppliers" providing "technology" which has specific "TOMs".

harsh: this seems relevant to procurement of technology which delaram is looking into

delaram: in addition to information, for AI the Users (AI Act entity) also need technical information and risk assessments to be shared with the technology

harsh: we can generalise these requirements as concepts applicable for technology e.g. documentation, instruction of use - which can be added to the TECH extension

group agrees that this would be an useful addition

ACTION: georg to share example of cloud technologies that Signatu wishes to model and the NIS2/DORA security requirements

ACTION: delaram to share concepts for what information is to be provided with the technology

GitHub discussions

Serialisation

<ghurlbot> Issue 118 Keep alt serialisations in the same folder (by coolharsh55)

There has been an ongoing discussion on GitHub regarding serialisation of DPV and how to serve the specific files and what filenames to use (technical discussion). FYI. This relates to keeping alternate serialisations in the same folder instead of having duplicates (e.g. RDFS+SKOS and OWL)

Domain/Range expression

<ghurlbot> Issue 119 Use dcam:domainIncludes and dcam:rangeIncludes (by coolharsh55) [concepts] [question]

Discussion on GitHub regarding use of dcam:domainIncludes and dcam:rangeIncludes to provide guidance on which concepts the properties can be used with. Suggestion to also include schema:domainIncludes and schema:rangeIncludes as commonly known semantics.

Updating concepts for non-personal data

<ghurlbot> Issue 99 Proposal to change DPV scope to include Non-Personal Data (by coolharsh55)

harsh: In preparation for expanding the scope of DPV to include or allow use with non-personal data, we need to change the definitions and descriptions of concepts to say "data" instead of "personal data". Volunteers for this are needed.

ACTION: georg and paul to go through the spreadsheet for non-personal data changes required

Updates to documentation

harsh: working on updating the repo structure and the way documentation is generated, have live version at https://harshp.com/dpv/dpv/

harsh: documentation is structured to be less dense - where a list of concepts is shown in the section instead of full tables to provide convenience. There is a full table of concepts at the end where definitions link to.

Categories of Data Subjects

<ghurlbot> Issue 116 Add Intended and Active Data Subject categories (by coolharsh55) [concepts] [question]

the group discussed the usefulness of concepts and that there are lots of possibilities for modelling entities and their specific aspects

the group agreed that it would be better to provide specific modelling of Active Data Subject in addition to any contextual ActiveInvolvement

the group agrees that providing isDeterminedBy would be useful

the group discussed Expected/Unexpected and considers the possible confusion arising out of this - so for now will keep this as a proposal

Automation and Human Involvement

<ghurlbot> Issue 108 Revise Automation and HumanInvolvement concepts (by coolharsh55)

delaram: confusion between concepts related to automation - unclear what AssistiveAutomation and PartialAutomation have as differences.

harsh: since these are ISO defined terms, we have to find suitable interpretations to keep them in the same structure, but yes we would need to resolve the confusion here

delaram: human involvement has differences from subjects and the concepts in the ISO document

harsh: yes, the human involvement concepts are a mixed set from different sources e.g. 22989, AI principles documents, existing DPV concepts

delaram: human in the loop and other concepts are non-normative

harsh: there are too many diverging uses of these with no consensus, so the proposed concepts are broadly categorised as in the loop (involved) with specific types of involvements or out of the loop (not involved)

Discussed without updates

Implementing 27560 using DPV

Modelling information on legal bases other than consent

Next Meeting

harsh: The next meeting will be in 1 week on WED OCT-25 15:00 WEST / 16:00 CEST. Agenda includes discussions on: 1) Active/Passive etc. concepts ; 2) Automation and Human Involvement concepts ; 3) AOB.

Summary of action items

  1. georg to share example of cloud technologies that Signatu wishes to model and the NIS2/DORA security requirements
  2. delaram to share concepts for what information is to be provided with the technology
  3. georg and paul to go through the spreadsheet for non-personal data changes required
Minutes manually created (not a transcript), formatted by scribe.perl version 217 (Fri Apr 7 17:23:01 2023 UTC).