Location
Hilton Hawaiian Village, Honolulu, Hawaii
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2024 12:00 AM
End Date
6-1-2024 12:00 AM
Description
Artificial Intelligence (AI) is becoming a crucial part of our lives. Although AI applications, such as facial recognition, autonomous driving and ChatGPT, can benefit different industries, users are more and more concerned about the ethical issues associated with AI systems. As a result, various ethics frameworks and standards have been proposed for regulating AI systems. Nevertheless, existing ethics frameworks and standards are hardly actionable or implementable for AI developers. To fill this gap, the current study proposes an actionable ethics-aware guideline for AI developers, as well as a set of quality metrics for ethical AI systems. Further, we implement the guideline using numerous AI predictive models constructed on a national big data set that estimates children’s risk of experiencing abuse and neglect in the United States. Evaluation results indicate that the proposed guideline can effectively enhance the quality of predictive models in utility, ethicality and cost dimensions.
Recommended Citation
Han, Yuzhang; Landau, Aviv; Kulkarni, Paritosh; Modaresnezhad, Minoo; and Nemati, Hamid, "An Implementable Guideline for Developing Ethical AI Systems: The Evaluation of Child Abuse and Neglect Prediction" (2024). Hawaii International Conference on System Sciences 2024 (HICSS-57). 2.
https://aisel.aisnet.org/hicss-57/sj/digital-discrimination/2
An Implementable Guideline for Developing Ethical AI Systems: The Evaluation of Child Abuse and Neglect Prediction
Hilton Hawaiian Village, Honolulu, Hawaii
Artificial Intelligence (AI) is becoming a crucial part of our lives. Although AI applications, such as facial recognition, autonomous driving and ChatGPT, can benefit different industries, users are more and more concerned about the ethical issues associated with AI systems. As a result, various ethics frameworks and standards have been proposed for regulating AI systems. Nevertheless, existing ethics frameworks and standards are hardly actionable or implementable for AI developers. To fill this gap, the current study proposes an actionable ethics-aware guideline for AI developers, as well as a set of quality metrics for ethical AI systems. Further, we implement the guideline using numerous AI predictive models constructed on a national big data set that estimates children’s risk of experiencing abuse and neglect in the United States. Evaluation results indicate that the proposed guideline can effectively enhance the quality of predictive models in utility, ethicality and cost dimensions.
https://aisel.aisnet.org/hicss-57/sj/digital-discrimination/2