The event is jointly organized by CWI and Maastricht University to strengthen collaborations regarding responsible AI across the country. The symposium is in the context of an extended research visit of Nava Tintarev to CWI. Read more about our visiting programme.
About the symposium
In the symposium on 'Understanding Bias in AI" we intend to adopt a broad, systemic lens that sees AI as a component within a complex sociotechnical system involving institutions, actors and processes. The lifecycle of an AI system is shaped by institutional and operational contexts which influence decisions at each step: 1) the selection and curation of training data, 2) the methodologies by which models are trained and validated, 3) how the system is deployed (e.g., on which instances, for which task etc. ), and finally, 4) the manner in which its outputs are communicated to and interpreted by users.
At each step of this lifecycle, bias may take root, propagate, or be inadvertently amplified. With this symposium, we hope to create a space for critical, cross-disciplinary reflection on how bias manifests across the AI lifecycle, and the interplay between human and machine biases.
