Acquisition gets data that fits the task, is representative of the population and includes the labels the task needs.
What the base framework establishes
The clause treats acquisition as the first pipeline stage. Training and evaluation both depend on the data, so it must suit the business purpose. Training data is often a subset of a larger population and should represent that population. Supervised tasks need labels or target variables, and the system also needs validation and test data, which can be acquired with the training data or separately. The type of data follows the problem: transactional data for forecasting, images for picture classification, and so on. Sources are identified and the data is acquired, from a store or a stream, and common sources include transactions, sensors, queries, images, documents, sound and social media.
Why it matters to an AI management system
Representativeness is the question that decides whether a result travels. A model trained on one region, one device or one customer segment will be used on another, and the framework points at that risk before any model exists. The acquisition record is also where licence, consent and provenance live.
The public draft summary names no change to 8.2 Data acquisition. The final text could still touch it.
Evidence a reviewer samples
| Category | Artefact |
|---|---|
| Acquisition record | Source, method, date range and the population the data represents. |
| Representativeness note | How the sample compares with the population it will serve. |
| Licence and consent record | The right to use the data for training and the basis for it. |
One question to prepare for. What population does this training set represent, and where will the model be used?
Common gaps
- The data is whatever was easiest to export, with no representativeness claim.
- Validation and test data are acquired later from a different population.
- The licence covers storage but not model training.
Prepare before publication
Add a representativeness note and a licence line to each acquisition record. Acquire the evaluation data with the training data where possible.
Dates that matter
No separate date attaches to 8.2; the record dates are the stage date, the planning estimate and the CEN expectation stated above.
Cross-references
3.3 Data · 6.4 Data · 8.3 Data preparation · 8.5 Verification and validation
Source
Source for 8.2: ISO/IEC 23053:2022, clause 8.2. The public sample stops before 8.2, so this page relies on the public contents and the public summary and does not quote the clause. The two official records are linked on this page so any line can be checked.
Not legal advice; does not certify. A reading of the framework and the evidence a reviewer commonly samples, not a substitute for the standard or your auditor's rules.