Bias Is Not Just In The Dataset
Identify systemic, statistical, and human sources of AI bias in a workplace AI use case.
Bias mitigation is source work before it is model work. The Three Sources NIST SP 1270 frames AI bias as socio-technical. Systemic bias comes from institutions, policies, access patterns, and historical inequities. Statistical bias comes from data, sampling, measurement, proxies, model objectives, and evaluation choices. Human bias comes from labelers, designers, reviewers, operators, and decision makers. The categories overlap, but separating them makes the work actionable. Why The Lens Works A single aggregate metric can hide very different failure modes. If the training data underrepresents a group, the remedy may be data collection and subgroup testing. If the target variable…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in