AI Governance Institute
All governance templates →Is our training data compliant with global privacy laws?

Implementation Kit

AI Training Data Provenance Template and Privacy Checklist

Whether you may use a dataset for this training purpose, and what you owe the people in it. A provenance record, a PII assessment method, a minimization checklist, and a data subject rights response template for training data.

Who this is for: The ML lead and privacy partner clearing a dataset before it is used to train or fine-tune a model.

Download the kit (Markdown) ↓4 artifacts. Every table also copies as CSV.

1. Training data provenance record

Spreadsheet

One per dataset. Source, the basis it was collected under, and whether training is a compatible use.

Template

FieldEntry
Dataset name and version
Source (internal system, vendor, public, scraped)
Original collection context and date range
Legal basis for original collection
Training-use compatibility assessmentcompatible / incompatible / needs new basis + reasoning
PII assessment resultnone / de-identified / contains PII (see PII assessment)
Licensing / terms of use for the data
Approved for training byname, date

Worked example

FieldEntry
Datasethiring-outcomes-2021-2026 v3
Sourceinternal HRIS
Original contextrecruitment administration; records 2021-01 to 2026-06
Legal basislegitimate interest (HR administration)
Training-use compatibilityNeeds assessment: model training is a new purpose. Compatibility memo concludes it is compatible given the recruitment nexus and anonymization; approved by DPO.
PII assessmentContains PII in raw form; training extract de-identified (see PII assessment)
Licensinginternal data, no third-party terms
Approved for training byDPO + ML Lead, 2026-07-05

Acceptance criteria

  • Every dataset used in training or fine-tuning has a provenance record.
  • The training-use compatibility assessment has explicit reasoning, not just a yes.
  • The record is retained as part of the model documentation.

2. PII assessment methodology

Spreadsheet

A consistent way to classify what identifiers a dataset contains.

Template

CategoryDefinitionExamplesHandling
Direct identifiersIdentify a person on their ownname, email, phone, government ID, account numberremove or tokenize before training
Quasi-identifiersIdentify in combinationpostcode, birth date, job title + employer, rare attributesgeneralize, bucket, or assess re-identification risk
Special-category dataSensitive under lawhealth, race, religion, sexual orientation, biometrics, trade unionremove unless a specific condition is met and documented
Free-text fieldsMay contain any of the aboveCV text, notes, transcriptsrun entity detection; redact or exclude

Worked example

Applied to hiring-outcomes-2021-2026:

FieldCategoryAction taken
candidate_name, emailDirectremoved; replaced with a salted hash not used as a feature
postcode + application_dateQuasipostcode truncated to district; date bucketed to month
CV free textFree-text (mixed)entity detection for names, contact info, and health mentions; redacted
self-identified ethnicitySpecial-categoryexcluded from the training extract entirely; retained separately for bias testing only
Re-identification risk on the extract assessed as low (k >= 20 on quasi-identifier combinations).

Acceptance criteria

  • Every field is placed in a category, and free-text fields are scanned, not assumed clean.
  • Special-category data is excluded from training inputs unless a documented condition applies.
  • Re-identification risk on the final extract is assessed, not just direct identifiers removed.

3. Data minimization checklist

Spreadsheet

Cut the dataset to what the training purpose actually needs.

Template

CheckDoneNote
Every field has a stated reason it is needed for this modelY / N
Fields with no modeling value removedY / N
Precision reduced where full precision is not needed (dates to month, location to region)Y / N
Records aggregated or sampled where individual rows are not requiredY / N
Retention limit set for the training extractY / N
Differential privacy or noise considered for sensitive aggregatesY / Ndecision + reason

Worked example

CheckDoneNote
Field-by-field justificationY41 source fields reduced to 18
No-value fields removedYdropped internal workflow IDs, recruiter names
Precision reducedYdates to month; location to district
Aggregated / sampledPartialfull rows needed for the label; no down-sampling
Retention limitYextract deleted 18 months after the model version retires
Differential privacyNdecided against for a supervised model on this scale; noted

Acceptance criteria

  • Each retained field has a documented modeling justification.
  • Precision and retention are reduced to what the purpose needs.
  • The minimization decisions are recorded in the provenance record.

4. Data subject rights response template (training data)

Document

A consistent answer when someone asks what happened to their data in your training set.

Template

Use for access, erasure, and objection requests that touch training data.

  • Was the requester's data in a training set? how you checked, and the result
  • Access response: what categories of their data were used, for which model versions, and the source dataset
  • Erasure response: removal from raw datasets and operational stores; position on the trained model (retrain cadence, filtering, or a documented feasibility limit)
  • Objection response: how future processing of their data is prevented (exclusion from the next training extract, suppression at inference)
  • Machine unlearning statement: your standing position and compensating measures
  • Timeframe and contact

Worked example

  • In a training set? Checked the hashed-identifier index for the 2021-2026 extract: yes, the requester's application record was included.
  • Access: categories used were work history, education, and application date (generalized); model versions 2.0 through 2.4; source dataset hiring-outcomes-2021-2026.
  • Erasure: raw record deleted from HRIS and the raw dataset; the training extract is de-identified with no recoverable link, so targeted removal from the trained model is not feasible; the next scheduled retrain (Q1) will be built from a rebuilt extract that excludes the deleted record.
  • Objection: the requester is added to the training exclusion list and an inference suppression list.
  • Unlearning statement: we do not perform per-record unlearning on deployed models; compensating measures are the exclusion list and the retrain cycle.
  • Timeframe: completed within the statutory window; contact privacy@company.example.

Acceptance criteria

  • You can actually determine whether a given person was in a training set.
  • The response distinguishes raw datasets from the trained model and is honest about feasibility.
  • Objection produces a concrete exclusion, not only an acknowledgement.

Governance controls this kit produces evidence for

Completing the artifacts above gives you a head start on the evidence requirements for these controls.

DGC-001
DGC-001

The provenance record is the training data provenance and lawful-basis documentation.

DGC-002
DGC-002

The PII assessment method is the PII handling standard applied to training data.

DGC-003
DGC-003

The minimization checklist is the data minimization record for training datasets.

CMP-007
CMP-007

The provenance and PII records supply the training-data governance evidence an EU AI Act high-risk conformity assessment requires.

ALC-004
ALC-004

Provenance records retained with the model support explainability and audit documentation.

This kit backs one playbook. Read the full guidance for the reasoning behind each artifact.

Decide what to implement next

Assess your governance gaps, then create an action plan with owners and target dates. Build and export without an account; sign in when you want to save your plan.

Start the AI governance assessment →