An inclusive algorithm is a decision-making system, whether statistical, rule-based, or learned, that is deliberately audited and adjusted so that its outputs do not systematically disadvantage women, girls, or other groups historically excluded from the data that trained it. Inclusion here is a property to be engineered and verified, not a byproduct hoped for.
Bias in an algorithm rarely announces itself. It surfaces as a lower loan approval rate, a resume filtered out before a human reads it, a health risk score calibrated on a population that does not include the patient in front of the clinician. Naming the algorithm as the site of intervention, rather than the individual outcome, is what allows the harm to be traced and corrected at its source.
The A+ Alliance approach: Inclusion is the organizing word in the Alliance’s own name. The A+ Alliance for Inclusive Algorithms is a global, multidisciplinary coalition, led by Women at the Table and Code for Africa, that brings academics, activists, and technologists together to prototype and audit automated decision-making systems for gender, racial, and intersectional bias.
Related projects
Our Ecosystem, A+ Advisory Board, Gender & AI Innovation Collective
Related resources
A+ Declaration, Global Directory, Feminist AI Network Papers
