The FATE (Fairness, Accountability, Transparency, and Ethics) group at MSR Montréal is a new team that seeks to help align the uses of computational systems to broader goals in the service of social good, justice, and fairness. Our aim is to understand the limits of computational systems and their impacts on individuals and societies. To these ends, our team studies how societal issues like systemic bias and discriminatory behaviors are amplified by and manifest in complex real-world sociotechnical systems. We develop both quantitative and qualitative techniques to help mitigate such issues and guide the design and application of sociotechnical systems.
We frequently engage in collaborative, multidisciplinary research with both internal and external partners, as well as foster close collaborations with product groups (Bing, MS Office, Xbox, Microsoft Turing), aiming to bring in a deeper domain and product grounded perspective to the broader Responsible AI research community’s agenda. Our use-inspired, mixed-methods approach to research recognizes that the hard-to-operationalize, complex nature of many FATE challenges requires interdisciplinary mitigation strategies.
Our current projects focus on:
- Responsible natural language generation (NLG) and issue discovery frameworks
- Objectionable behaviors, toxicity, and addiction in sociotechnical systems
- Harms in information systems and other multi-stakeholder environments