Quality of Data for AI Systems
Organisations must define data quality standards and ensure their AI systems' data complies with these standards.
Plain language
Ensuring data quality means checking that the information our AI uses is accurate and reliable. Imagine if your sales tool recommends the wrong products because the data is outdated; maintaining high-quality data prevents such costly mistakes.
Framework
ISO/IEC 42001:2023
Control effect
Preventative
Classifications
N/A
Official last update
01 Dec 2023
Control Stack last updated
19 May 2026
Maturity levels
N/A
Official control statement
The organisation shall define and document requirements for data quality and ensure that data used to develop and operate the AI system meet those requirements.
Why it matters
If your AI uses bad data, it might give customers wrong information, which could harm your business's reputation and result in financial losses.
Operational notes
Review the data quality checklist every time new data is added to ensure it still meets the criteria, not just during periodic audits.
Implementation tips
- The data steward should create clear guidelines outlining what 'good quality' data looks like. This might involve regular spot-checks where you compare sample AI outputs with expected results to catch errors early.
- AI leads should collaborate with data providers to ensure ongoing data quality. They can set up periodic reviews to verify the data remains accurate, using a simple checklist of criteria that matter most for your business outcomes.
- CISOs should implement security measures that protect data integrity. Use encryption for data storage and transfer, ensuring no one can tamper with the data the AI system uses.
- The product owner should regularly review AI outputs for unexpected results that suggest poor data quality. Keep a simple log of anomalies, and discuss them in monthly team meetings to address any data quality issues quickly.
- Procurement should include clauses in supplier contracts specifying data quality expectations. Agree on specific data quality metrics and require regular data quality reports from suppliers.
Audit / evidence tips
- AskRequest the organisation's documented data quality requirements. GoodThe document clearly defines data quality standards, is accessible to staff, and is updated within the last year.
- AskRequest the latest data quality review report. GoodThe report is dated within the last six months, highlights issues found, and outlines actions to be taken.
- AskRequest recent training data samples. GoodSamples align with the defined data quality standards in every aspect.
- AskLook at the supplier contract for data quality clauses. GoodThe contract specifies quantifiable data quality metrics and reporting requirements.
- AskRequest the incident log for AI data issues. GoodEach logged incident includes an action plan and follow-up within a month.
Cross-framework mappings
How Annex A 7.4 relates to controls across ISO/IEC 27001, ISO/IEC 42001, Essential Eight, and ASD ISM.
ISO 27001
| Control | Notes | Details |
|---|---|---|
| handshake Supports (2) expand_less | ||
| Annex A 8.9 | Annex A 7.4 requires defined and documented data quality requirements and assurance that AI training/operational data meets them | |
| Annex A 8.32 | Annex A 7.4 requires the organisation to define, document, and assure data quality requirements for data used to develop and operate AI s... | |
ASD ISM
| Control | Notes | Details |
|---|---|---|
| handshake Supports (1) expand_less | ||
| ISM-0042 | Annex A 7.4 requires the organisation to specify data quality requirements and ensure AI system data conforms to them | |
These mappings show relationships between controls across frameworks. They do not imply full equivalence or certification.
Want to implement this AI control?
Mindset Cyber runs PECB-accredited ISO/IEC 42001 training that maps directly to the AI controls in this library.