Abstract

The dataset contains the matching of job positions and hiring candidates; this data has been collected by a big Italian company, working in the HR sector - ADECCO. The detailed description of the data can be found in a README file within the compressed archive. This activity is part of the HORIZON-CL4-2021-HUMAN-01-24-AEQUITAS project (g.a. 101070363). The aim of AEQUITAS to address and tackle the multiple manifestations of bias and unfairness in Artificial Intelligence (AI) from a variety of dimensions, such as the development of AI tools, the data used to train, test and validate them or the interpretation practices developed around them. AEQUITAS offers an overall approach for tackling the problem, savant of the criticalities that automation and AI techniques bring about. This case study allows for detecting hiring dataset biases which are the primary source for training a novel AI system. For example, there are historical trends in the labour market in favour of men (higher levels of education, once of employment, hence of hiring) that might be reflected in the data history of ADECCO. Balancing these inequalities in data or leaving data biased and targeting debiasing or bias reducing algorithms is a key step for a fair AI system and dataset. The ADECCO data offer data were to compare selection decisions with regards to important bias, such as gender, age, economic background, etc.