AI and gender pay gap

Important results from the project

The project tested whether AI can help manage gender bias in salary setting by developing a prototype that can identify patterns in salaries and indicate unfair differences for equal work in the banking sector. The AI prototype can be used to analyze salary levels so that women and men who perform equal work, have similar levels of responsibility, education and experience receive compensation that does not differentiate between the sexes. The AI prototype can provide a better understanding of where pay gaps exist and can indicate where action is needed and a Roadmap has been developed.

Expected long term effects

The project has resulted in an AI prototype that can assist changes in recruitment and salary setting and has potential to stimulate so that women who previously had a disadvantageous position can obtain better conditions for equal compensation. Development of machine learning technology enables learning for solutions in other contexts to counteract gender bias in development and application. Roadmap and algorithm development have been reported in research report. The project has conducted workshops with actors interested in equal recruitment and wage setting.

Approach and implementation

Using norm criticism and algorithms as a method, the project explored how innovation in AI can reduce unconscious gender bias in recruitment and salary setting. The project was based on a scientific, systematic co-creative norm-critical approach where experimentation and evaluation of data was used in the development of the algorithm. Iterative norm-critical work was the basis for developing the algorithm where stereotypical gender norms could be avoided. Workshop meetings have been held continuously and resulted in the building of a database and development of the algorithm.