DPVCG Meeting Call

Repository: w3c/dpv

Agenda: https://www.w3.org/events/meetings/178d1c71-a92d-4da7-a196-6a89d0fe2277/20250529T133000/

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

Persistent ID for current minutes: https://w3id.org/dpv/meetings/meeting-2025-05-29

AOB

GeorgKrog: identified concepts from EHDS and AI Act for DPV, with EHDS concepts in spreadsheet (shared) and AI Act concepts to be added to spreadsheet

ACTION: Harsh to review EHDS concepts, Stratis to also review

HarshPandit: For the AI Act, we have existing concepts from Delaram's work on AIRO and VAIR which have been integrated into the AI and AI Act DPV extensions. There is more work from the TAIR project https://tair.adaptcentre.ie/ which would be helpful to avoid duplication and to use resources developed by DPVCG members.

ACTION: Julio and Delaram to share work done in TAIR with Georg and Harsh, and then consolidate with Georg's proposal for AI Act so we don't repeat work

GeorgKrog: Sebastian Hellensleben shared work in CEN/CENELEC on standards for AI Act https://www.linkedin.com/posts/sebastianhallensleben_jtc21-status-dashboard-activity-7331425771237826560--xX8 – would be interesting for the group to map these

GeorgKrog: (continued) https://www.linkedin.com/posts/activity-7332406233150164993-NfOw link to standards in detail – see vocabulary

HarshPandit: standards – 22989 same (EN), look at concepts and whether they are in scope, how to align with DPV, etc. lifecycle ones should be of interest to look at

GeorgKrog: are we modelling all of these?

HarshPandit: we have some of these, but others we need to think of how to model as they sometimes have multiple uses and definitions

StratisKoulierakis: to what degree are we looking for conformity with the AI Act? e.g. the word intended for intended purpose in Art.6, whereas we also have "intended to be used as safety component"

HarshPandit: as much as possible as we have the AI Act where such terms can be defined, so if something is defined in the AI Act then we keep it as phrased in the act as much as we can

Representing Use-Cases

<ghurlbot> Issue 284 [NEW]: Modelling Intended Uses/Purposes as Use-Cases (by coolharsh55)

continued discussion

discussing https://harshp.com/dev/dpv/use-cases-01

HarshPandit: we have dpv:UseCase and dpv:hasUseCase (we have agreement on these in principle) and we are now discussing hierarchy of potential categories of use-cases with dpv:UseCase as parent. Question: what categories of use-cases should we additionally define?

below is outcome of discussion and voting using document

deontic / normative categorisations -- acceptable, permitted, recommended, deterred, and prohibited have 3 votes

Acceptable 👍 3 votes 👎 no votes; Permitted (MAY) 👍 3 votes 👎 no votes; Recommended (SHOULD) 👍 3 votes 👎 no votes; Required (MUST) 👍 2 votes 👎 no votes; Tolerated (MAY NOT) 👍 one vote 👎 no votes; Unacceptable 👍 2 votes 👎 no votes ; Deterred (SHOULD NOT) 👍 3 votes 👎 no votes; Prohibited (MUST NOT) 👍 3 votes 👎 no votes

teleological -- intended and unintended have 3 votes

Intended (deontically Acceptable) 👍 3 votes 👎 no votes; Unintended 👍 3 votes 👎 no votes

Question: how to model Intended Purpose in AI Act with the above? -- extending a core concept in DPV is preferred with 3 votes

by extending dpv:UseCase 👍 2 votes 👎 no votes; by extending dpv:IntendedUseCase 👍 3 votes 👎 no votes

Next week discuss these most voted concepts

Operating Factors / Environment

<ghurlbot> Issue 285 [NEW]: Modelling Operating Factors for Technology (by coolharsh55)

HarshPandit: Why are we modelling these? Laws, like GDPR, EHDS, and Cybersecurity Act require description of how the technology functions. However, these are yet to be explained clearly in terms of guidance and case law. Therefore, we are trying to model things in preparation for eventual clarity.

Discussion on scope of these concepts: Let's discuss this, but keep these as proposed 👍 3 votes 👎 no votes; Let's add these now, revise later 👍 no votes 👎 no votes

Question: Do we provide tech:OperatingEnvironment? 👍 4 votes 👎 no votes ; defined as the environment (physical and virtual) within which the technology operates or will operate. This includes possibly everything that could be relevant, including location, data, people, machinery (other tech), etc.

Do we model more specific environments like physical, virtual, how technology is set up, how it is placed in relation to surroundings? 👍 no votes 👎 3 votes

Do we provide tech:OperatingFactor? 👍 4 votes 👎 no votes ; defined as a factor which influences the operation of the technology, and also as parts of the technology and of the environment which affect the technology i.e. they are a "factor" in relation to the technology's operations

Do we model more specific categories of operating factors? 👍 2 votes 👎 no votes

If yes, which categories? (shown in hierarchy)

HarshPandit: Requirements oriented: Required 👍 3 votes 👎 no votes ; Optional 👍 3 votes 👎 no votes ;

Effect oriented: Suitable 👍 2 votes 👎 no votes (is okay to have); Ideal (gives optimal results) 👍 2 votes 👎 no votes (may involve optional factors as well as necessary factors); Unsuitable 👍 2 votes 👎 no votes (is not okay to have); Problematic (gives problematic results) 👍 2 votes 👎 no votes ; Untested 👍 2 votes 👎 no votes

HarshPandit: Relation between factors and existing concepts like data, processing, technology, etc. https://harshp.com/dev/dpv/tech-operating-factors#Sec3 gives a list of existing properties in DPV and how they relate to the Factor concept. idea: we provide categories of factors (as above) to describe their relation to the operation i.e. required, suitable etc. and then the existing properties describe the factor in more detail e.g. tech has required factor as specific categories of data; e.g. tech has ideal factor – where the factor is described using 1) specific categories of data 2) from specific sources and 2) which has quality assurance (org measure). In the above, we are still describing "data" as a factor, as in it is a single thing, and therefore this is distinct from an use-case

Conclusions: We see value in providing information about factors and how they relate to the technology and where / whether they are suitable, but do not yet have clarity on how these should be modelled or how these will work with other concepts.

Secondary Use

<ghurlbot> Issue 283 [NEW]: Secondary Uses for Data Reuse (by coolharsh55)

continued discussion

discussing https://harshp.com/dev/dpv/secondary-uses

Question: what should we call the top concept for this? dpv:CompatibilityOfUse 👍 2 votes 👎 no votes (has "of" which we avoid if we can); dpv:Compatibility 👍 2 votes 👎 no votes (will need compatible/incompatible concepts); dpv:ReuseCompatibility 👍 3 votes 👎 no votes (can imply only reuse/secondary) -- most preferred option

Question: what concepts should we model (shown in hierarchy)? dpv:PrimaryUse 👍 3 votes 👎 no votes ; dpv:PrimaryInitialUse 👍 3 votes 👎 no votes ; dpv:PrimaryCompatibleReuse 👍 3 votes 👎 no votes ; dpv:SecondaryUse (incompatible with primary) 👍 3 votes 👎 no votes ; dpv:SecondaryContextualReuse 👍 2 votes 👎 no votes ; dpv:TertiaryUse (secondary + different context) 👍 2 votes 👎 no votes -- 3+ votes are accepted for continued discussion, rest need to be clarified

Next Meeting

The next meeting will be on JUN-05 Thursday 13:30WET/14:30CET

Agenda will be EHDS concepts led by Georg, and continued discussions from today on use-case categories, operating factors, and secondary use concepts