ASD ISM 2088 Ensure accuracy of AI model training data
What it says
Data validation and verification techniques are used to ensure the reliability and accuracy of training data used by AI models.
Quoted as published. Everything else on this page is written by Control Stack.
What it means
In plain EnglishTechniques verify that AI training data is reliable and accurate.
Ensuring the data used to train AI is accurate matters because if the training data is faulty, the AI could make poor decisions leading to bad outcomes like incorrect forecasts or flawed insights that affect your business success.
Free. Save your status, tasks and notes.
Why it matters
Inaccurate AI training data can lead to poor business decisions, costing time, money, and reputation for errors that could have been avoided.
How to implement
Data managers should verify the source of the training data.
They can do this by checking the data origin, ensuring it's from a trusted provider, and confirming that the data hasn't been tampered with since collection.
The IT team should implement regular data audits.
This involves setting regular intervals to check the data for errors and inconsistencies, which helps ensure the data is clean and usable.
Business analysts should define clear data accuracy criteria.
They can do this by stating what constitutes 'accurate data' for their specific needs, like checking for complete records and valid entries.
The staff responsible for data entry should ensure the data quality.
This can be done by validating the entered data against known standards at the point of entry to catch errors immediately.
System owners should facilitate training programs for staff who handle data.
These programs should cover their role in maintaining data accuracy and how to detect unreliable data early.
Keep it running
Regularly verify and update training data sources to maintain AI model accuracy and usefulness.
Evidence and audit
- Check 1 of 5What an auditor asks for
Ask for
data sourcing records: Request the documentation that outlines where the training data was sourced from.
Look at
evidence of trusted sources and confirmation of data integrity.
Good evidence
shows records of trusted, verified origins with timestamps.
- Check 2 of 5What an auditor asks for
Look at
reports that detail regular checks on data accuracy and integrity. Verify reports include identified and corrected errors.
Good evidence
report lists specific inaccuracies found and actions taken to correct them.
- Check 3 of 5What an auditor asks for
Ask for
data validation protocols: Obtain documentation on procedures used to ensure data accuracy. Check these protocols specify step-by-step verification processes.
Good evidence
protocol includes clear validation processes and staff responsibilities.
- Check 4 of 5What an auditor asks for
Look at
records of training programs focused on data accuracy. Ensure these sessions are regular and include all relevant personnel.
Good evidence
includes training dates, attendees, and content covered.
- Check 5 of 5What an auditor asks for
Ask for
error logging and resolution records: Review records of data errors and how they were resolved. Ensure these logs detail the type of error and corrective action taken.
Good evidence
record shows timely error detection and resolution.
Mappings
Closest match in ISO/IEC 42001: A.7.2, Data for development and enhancement of AI system, which partially overlaps with this control.
ISO/IEC 42001: 5 mappings
This control partially overlaps A.7.2.
Annex A 7.2 requires the organisation to define, document and implement end-to-end data management processes for developing and enhancing AI systems. ISM-2088 relates by requiring validation and verification techniques to ensure the reliability and accuracy of AI model training data, which is a core activity within those data management processes. The overlap is specifically around data quality controls for training datasets, while Annex A 7.2 is broader and also covers governance, lineage, handling, retention and ongoing dataset management beyond accuracy checks.
This control partially overlaps A.7.3.
Annex A 7.3 requires the organisation to document the acquisition sources and selection criteria for data used in AI systems. This relates to ISM-2088 because data validation and verification to ensure reliability and accuracy are often part of the documented selection criteria and acceptance checks for training datasets. ISM-2088 focuses on technical validation quality controls for training data, while Annex A 7.3 is a governance requirement to record acquisition and selection details that apply beyond training data.
This control partially overlaps A.7.4.
Annex A 7.4 requires the organisation to define, document and verify data quality requirements for data used to develop and operate AI systems, emphasising dimensions such as accuracy and completeness. ISM-2088 focuses specifically on using data validation techniques to ensure the reliability and accuracy of AI model training data, which aligns with the quality verification activities in Annex A 7.4. While Annex A 7.4 is broader, encompassing various quality dimensions and applications beyond training data, the overlap occurs in the shared goal of ensuring data reliability.
This control partially overlaps A.7.6.
Annex A 7.6 requires the organisation to define and document criteria for selecting data preparation steps and methods so data is prepared consistently for AI use. ISM-2088 relates by requiring validation and verification techniques to ensure AI training data is reliable and accurate, which are common data-preparation activities that should be selected and documented under the Annex A 7.6 criteria. ISM-2088 is outcome-focused on accuracy/reliability, while Annex A 7.6 is governance-focused on how preparation methods are chosen and standardised across use cases.
This control depends on A.7.5.
Annex A 7.5 requires a process for recording data provenance over the life cycles of the data and AI system. ISM-2088 supports this by requiring validation and verification techniques to ensure AI model training data's reliability and accuracy, thus contributing to the trustworthiness of provenance records. ISM-2088 focuses on data quality, while Annex A 7.5 focuses on traceability and documentation.
Mappings show how controls relate across frameworks. They do not imply full equivalence or certification.