Black box algorithms and the rights of individuals: no easy solution to the “explainability” problem
Jarek Gryz, Marcin Rojszczak
Abstract
The design of modern machine learning systems should take into account not only their effectiveness in solving a given problem, but also their impact on the rights of individuals. Implementing this goal may involve applying technical solutions proven in the IT industry, such as event logs or certification frameworks.
Topics & Concepts
CertificationBlack boxComputer scienceEvent (particle physics)AlgorithmMachine learningArtificial intelligenceLawPolitical scienceQuantum mechanicsPhysicsExplainable Artificial Intelligence (XAI)Ethics and Social Impacts of AIAdversarial Robustness in Machine Learning