Artificial intelligence is increasingly shaping decisions that affect people’s health, education, access to information and interactions with public institutions. As those systems become more powerful, researchers say preventing harm cannot be an afterthought.
Kaushik Dutta
Kaushik Dutta, professor of information systems at the University of South Florida, is helping define what responsible AI should look like as the technology moves beyond today’s generative AI boom.
In the editorial “Responsible AI and Data Science for Social Good,” submitted to the INFORMS Journal on Computing, Dutta and his co-authors examine responsible AI across six areas: judicial systems, education, communication, health care, bias and fairness, and interpretability.
Their central argument is that principles such as fairness, privacy, transparency and robustness should be incorporated throughout the life of an AI system — from its initial design to deployment and ongoing governance — rather than addressed separately after problems emerge.
The editorial also looks ahead to challenges in the post-generative AI era. Among the priorities are reducing bias in large language models, making AI-supported decisions more reliable and understanding how automated systems affect people and institutions over time.
Authors: Kaushik Dutta, University of South Florida; Ajay Kumar, EMLYON Business School; Zhiling Guo, University of North Texas; Martin Bichler, Technical University of Munich; Dursun Delen, Oklahoma State University; Ram Ramesh, University of Buffalo; Paul Brooks, Virginia Commonwealth University.