Researchers Eija Ahonen and Eeva Farén from the Master’s Degree Programme in Digital Business Management at Lapland University of Applied Sciences have published a master’s thesis exploring gender biases in AI systems and the effectiveness of mitigation strategies.

This comprehensive overview delves into the persistent issue of gender inequality in the digital age and offers consolidated mapping of the latest methods and current mitigation strategies to promote fairness. As AI becomes increasingly integrated into everyday technology, from recruitment software to language processing tools, the risks of embedded gender biases grow. These biases not only perpetuate existing inequalities but also pose significant ethical challenges.

The research conducted by Ahonen and Farén systematically reviews existing literature to identify and evaluate various strategies aimed at addressing and mitigating these biases. The thesis employs a comprehensive systematic literature review, gathering data from multiple studies to analyse the underlying mechanisms of gender bias within digital technologies. This method ensures a robust analysis of the effectiveness of current mitigation strategies.

Key findings

  • Gender biases in AI are prevalent across various applications, affecting outcomes in sectors such as employment, healthcare, legislation and education.
  • Effective mitigation requires a multifaceted approach, including algorithm adjustment, data set balancing, and the implementation of ethical guidelines.
  • Current mitigation practices show promise in reducing biases but require broader implementation and continuous development.


The thesis highlights the need for ongoing research into AI fairness, suggesting that future work should focus on developing more sophisticated tools and methods to detect and counteract biases before they affect AI outputs.

– Addressing gender bias in AI is not just about correcting inequalities; it’s about ensuring the technology we depend on reflects the ethical standards of our society, says Eija Ahonen.
– Our research indicates a promising direction towards achieving fairness in AI, but it also shows that much work remains to be done, adds Eeva Farén.


Conclusions

This thesis contributes significantly to the understanding of gender biases in AI, offering a broad overview of mitigation strategies, frameworks, techniques and tools on various industries, covering the research of past 2 years, leading to analysis of the complexity, challenges and significance of this field. It aims to foster a more equitable technological future and underscores the importance of integrating ethical considerations into AI development.

Contact

Eeva Farén eeva.faren(at)gmail.com
Eija Ahonen eija_ahonen(at)hotmail.com