1 - 4 of 4
ItemGuide for Artificial Intelligence Ethical Requirements Elicitation - RE4AI Ethical Guide( 2022-01-04)Development and use of Artificial Intelligence (AI) based systems are growing at a fast pace in our society, simultaneously ethical concerns are arising from them. Addressing AI ethics is a continual issue and has provoked much debate among researchers. The aim of this work is to provide a Guide for Artificial Intelligence Ethical Requirements Elicitation (RE4AI Ethical Guide). The Design Science Research methodology was adopted in order to understand the problem, develop a prototype and evaluate it through a survey. The proposed Guide, composed of 26 cards along 11 ethical principles, is both useful and practical and can help in the elicitation of ethical requirements for AI in the context of agile development. Our preliminary results reveal that the Guide contributes to bridging the gap between high-level and abstract principles and practice by assisting developers and Product Owners to elicit ethical requirements and implement ethics in AI.
ItemA Value Sensitive Design Perspective on AI Biases( 2022-01-04)Artificial Intelligence (AI) technology has made profound impacts in our society but concerns about AI biases are rising. This paper classifies AI-related biases and proposes strategies to tackle them. To inform our study, we review AI research on human values to identify three categories of AI biases: pre-existing, technical, and emergent. Informed by the value sensitive design (VSD) framework, we then map the AI biases to the three phases (conceptual, empirical, and technical) of VSD investigation. Our analysis shows that both conceptual and empirical investigations are helpful for addressing pre-existing bias, technical investigation for technical bias, and both technical and empirical investigations for emerging bias. The paper highlights that to effectively tackle AI-related biases, it is important for AI developers and the user community to understand human values in an AI context and to advocate for developing AI-specific value-oriented standards that are agreed upon and adopted by all stakeholders.
ItemAn Exploratory Study on Fairness-Aware Design Decision-Making( 2022-01-04)With advances in machine learning (ML) and big data analytics, data-driven predictive models play an essential role in supporting a wide range of simple and complex decision-making processes. However, historical data embedded with unfairness may unintentionally reinforce discrimination towards minority groups when using data-driven decision-support technologies. In this paper, we quantify unfairness and analyze its impact in the context of data-driven engineering design using the Adult Income dataset. First, we introduce a fairness-aware design concept. Subsequently, we introduce standard definitions and statistical measures of fairness to the engineering design research. Then, we use the outcomes from two supervised ML models, Logistic Regression and CatBoost classifiers, to conduct the Disparate Impact and fair-test analyses to quantify any unfairness present in the data and decision outcomes. Based on the results, we highlight the importance of considering fairness in product design and marketing, and the consequences, if there is a loss of fairness.