Unveiling Bias: GRASE Partners pave the way for a bias-free recruitment based on AI

Thanks to EU Funds, Skills Partners collective efforts are contributing to a more equitable and unbiased future 

The GRASE project takes a pioneering step forward with its focus on eradicating gender and race stereotypes in labor market access. This collaborative initiative involves prominent partners, with the Institute for the Study of Migration (ISMU) in Milano spearheading the charge. Spanning a two-year timeline from 2021 to 2022, this venture is financially supported by the European Union's Rights, Equality & Citizenship program.

At its core, GRASE addresses the dual discrimination faced by migrant women in their pursuit of employment, adopting a truly intersectional approach. The primary objective of the project is to facilitate the access of women with migratory backgrounds to the labor market. To achieve this, GRASE focuses on dismantling barriers encountered in career counseling services systems.  

An innovative toolkit to address gender and races biases in AI  

One notable innovation emerging from the project is an AI-based toolkit designed to identify gender and race biases concealed within the recruitment process. Given the increasing reliance on AI algorithms in recruitment, the need to scrutinize biases embedded in machine learning processes is more pertinent than ever. The GRASE toolkit operates on a lexicon of distorted terms specifically related to gender, enabling its adaptation for the recognition of other bias types.

The toolkit operates under two distinct logics. The first, a top-down approach, involves human intervention in the classification of terms. Individuals curate the lexicon by selecting terms relevant to gender biases. The second, termed bottom-up, utilizes artificial intelligence models to autonomously detect new biased words, adding a dynamic and self-improving dimension to the toolkit. What makes the GRASE toolkit stand out is its inherent flexibility. While originally designed to identify gender and race biases, its structure allows for seamless application in identifying other forms of bias. This versatility positions GRASE as a comprehensive solution in the ongoing battle against discrimination in various contexts beyond the labour market.  

GRASE Project Partners’ collective efforts are contributing to a more equitable and unbiased future. By addressing the nuances of intersectionality and harnessing the power of AI, GRASE paves the way for a workforce that is not only diverse but free from the shackles of gender and race biases.