SFS E / SF 0

Back to SFS E

Status: Question model review passed; solution and completed SF authoring pending.

This is a generated authoring draft. Human approval and operational assessment results have not been supplied.

Question

Did the organization maintain documented evidence that all AI models underwent bias evaluation using representative demographic datasets, with mitigation measures applied where necessary, throughout the review period?

Plain-language explanation

Model-reviewed question draft. Completed explanations and a canonical solution are pending. The proposed positive criterion is: Documented evidence exists showing bias testing was performed using representative datasets and mitigations were implemented where bias exceeded acceptable thresholds

Assessment criterion and evidence

Actor: Organization developing or deploying AI systems. Scope: All AI models trained and deployed during the review period. Evidence period: Through the entire review period, from model development through deployment. Required evidence: Records of bias testing methodologies and demographic representation metrics; Documentation of mitigation measures applied to training data or model outputs. Negative criterion: No documented evidence of bias evaluation or ineffective mitigation measures against identified demographic biases. Missing evidence: Pending until documented evidence of bias evaluation methodology and mitigation outcomes is provided

View the JSON record (opens in a new tab).

Assessment result: unset.