Fairness: How AI Creates and Maintains Inequalities
AI is increasingly used in highly sensitive areas such as medical diagnosis, human resource selection, or criminal justice. Therefore, the attention of professionals, users, and regulators is increasingly focused on the quality of decisions adopted, including the concern to guaranteeing its equitable, non-discriminatory nature.
Inequity can arise in AI through social bias built into training data sets, in decisions made during the machine learning process, and in the complex feedback loops that arise when a machine learning model is implemented in the real world.
In this talk, Josep will speak about the problem of the absence of Fairness in AI systems. The talk will be about some of the ways and tools we have today to measure and mitigate the existence of bias and ensure better decision-making.