Historical Bias
Bias that exists in real-world historical data and is encoded into AI training datasets, where past societal inequities are reflected and potentially amplified by the model.
In plain language
When unfairness from the past is baked into your training data. If historical hiring records show men were promoted more often, an AI trained on that data will learn to prefer men for promotions.
Why this matters
Historical bias is critical for AI governance because it means even accurate historical data can produce unfair outcomes. Your governance framework must require assessment of whether training data encodes past inequities and include mitigation strategies for high-risk applications.
Relevance
GovernanceAssessing historical bias in training data is necessary to identify and mitigate discrimination risk before AI systems are deployed.
Related terms
Putting historical bias into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.