Context: The increasing integration of artificial intelligence and machine learning into software systems has highlighted the critical importance of ensuring fairness in these technologies. Bias in software can lead to inequitable outcomes, making fairness testing essential. However, the current landscape of fairness testing tools remains underexplored, particularly regarding their practical applicability and usability for software development practitioners. Goal: This study aimed to evaluate the practical applicability of existing fairness testing tools for software development practitioners, assessing their usability, documentation, and overall effectiveness in real-world industry settings. Method: We identified 41 fairness testing tools from the literature and conducted a heuristic evaluation and documentary analysis of their installation processes, user interfaces, supporting documentation, and update frequencies. Technical analysis included assessing configurability for diverse datasets. The analysis focused on identifying strengths and deficiencies to determine their suitability for industry use. Findings: Our findings revealed that most fairness testing tools show significant deficiencies, particularly in user-friendliness, detailed documentation, and configurability. These limitations restrict their practical use in industry settings. The tools also lack regular updates and possess a narrow focus on specific datasets, which constrains their versatility and scalability. Despite some strengths, such as cost-effectiveness and compatibility with several environments, the overall landscape of fairness testing tools requires substantial improvements to meet industry needs. Conclusion: There is a pressing need to develop fairness testing tools that align more closely with industry requirements, offering enhanced usability, comprehensive documentation, and greater configurability to effectively support software development practitioners.

Context: This study explores how software professionals identify and address biases in AI systems within the software industry, focusing on practical knowledge and real-world applications. Goal: We focused on understanding the strategies employed by practitioners to manage bias and their implications for fairness debt. Method: We employed a qualitative research method, gathering insights from industry professionals through interviews and using thematic analysis to explore the collected data. Findings: Professionals identify biases through discrepancies in model outputs, demographic inconsistencies, and training data issues. They address these biases using strategies such as enhanced data management, model adjustments, crisis management, improving team diversity, and ethical analysis. Conclusion: Our paper presents initial evidence on addressing fairness debt and lays the groundwork for developing structured guidelines to manage fairness-related issues in AI systems.

ABSTRACT AI systems for decision‐making have become increasingly popular in several areas. However, it is possible to identify biased decisions in many applications, which have become a concern for the computer science, artificial intelligence, and law communities. Therefore, researchers are proposing solutions to mitigate bias and discrimination among decision‐makers. Some explored strategies are based on GANs to generate fair data. Others are based on adversarial learning to achieve fairness by encoding fairness constraints through an adversarial model. Moreover, it is usual for each proposal to assess its model with a specific metric, making comparing current approaches a complex task. Therefore, this work proposes a systematical benchmark procedure to assess the fair machine learning models. The proposed procedure comprises a fairness‐utility trade‐off metric (), the utility and fairness metrics to compose this assessment, the used datasets and preparation, and the statistical test. A previous work presents some of these definitions. The present work enriches the procedure by increasing the applied datasets and statistical guarantees when comparing the models' results. We performed this benchmark evaluation for the non‐generative adversarial models, analyzing the literature models from the same metric perspective. This assessment could not indicate a single model which better performs for all datasets. However, we built an understanding of how each model performs on each dataset with statistical confidence.

Context. Algorithmic racism is the term used to describe the behavior of technological solutions that constrains users based on their ethnicity. Lately, various data-driven software systems have been reported to discriminate against Black people, either for the use of biased data sets or due to the prejudice propagated by software professionals in their code. As a result, Black people are experiencing disadvantages in accessing technology-based services, such as housing, banking, and law enforcement. Goal. This study aims to explore algorithmic racism from the perspective of software professionals. Method. A survey questionnaire was applied to explore the understanding of software practitioners on algorithmic racism, and data analysis was conducted using descriptive statistics and coding techniques. Results. We obtained answers from a sample of 73 software professionals discussing their understanding and perspectives on algorithmic racism in software development. Our results demonstrate that the effects of algorithmic racism are well-known among practitioners. However, there is no consensus on how the problem can be effectively addressed in software engineering. In this paper, some solutions to the problem are proposed based on the professionals' narratives. Conclusion. Combining technical and social strategies, including training on structural racism for software professionals, is the most promising way to address the algorithmic racism problem and its effects on the software solutions delivered to our society.