BackInaccessible training data
Inaccessible training data
Risk Domain
Challenges in understanding or explaining the decision-making processes of AI systems, which can lead to mistrust, difficulty in enforcing compliance standards or holding relevant actors accountable for harms, and the inability to identify and correct errors.
"Without access to the training data, the types of explanations a model can provide are limited and more likely to be incorrect."
Entity: Who or what caused the harm
Intent: Whether the harm was intentional or accidental
Timing: Whether the risk is pre- or post-deployment
Supporting Evidence (1)
1.
"Low quality explanations without source data make it difficult for users, model validators, and auditors to understand and trust the model."
Other risks from IBM (2025) (63)
Lack of training data transparency
6.5 Governance failureHumanUnintentionalPre-deployment
Uncertain data provenance
6.5 Governance failureHumanOtherPre-deployment
Data usage restrictions
7.3 Lack of capability or robustnessHumanUnintentionalPre-deployment
Data acquisition restrictions
7.3 Lack of capability or robustnessHumanUnintentionalPre-deployment
Data transfer restrictions
7.3 Lack of capability or robustnessHumanUnintentionalPre-deployment
Personal information in data
2.1 Compromise of privacy by leaking or correctly inferring sensitive informationAI systemUnintentionalPost-deployment