Artificial intelligence (AI) applied to people management (selection, hiring, promotion, and remuneration) can perpetuate and amplify gender biases if algorithms learn from historical data with inequalities. In application of European and Spanish regulations, organizations can reduce this risk by auditing their AI systems, integrating the gender perspective, guaranteeing human supervision and training their teams to ensure that technology does not compromise equality in the workplace.
Every March 8, the date on which International Women’s Day is commemorated, it invites us to reflect on the progress made in terms of equality and analyze the new challenges that, if they are not already a reality, will be in the near future.
In the field of work and people management, one of the challenges in recent years comes from the growing use of artificial intelligence (AI) systems in the most essential processes of people management: selection, hiring, promotion or remuneration.
It is undeniable that technology and, specifically, AI offers efficiency, but also objectivity? It might seem complicated that an algorithmic system is not objective, since it precisely helps to eliminate or mitigate that intrinsic element to people that are unconscious biases. Now, what happens if we replace the human component —and therefore its biases— with an algorithm that has learned from historical data? What if the profiles of success in the organization have historically been male? What if the promotion system designed penalizes non-linear trajectories or with interruptions, for example, such as those produced by care?
The answer is that learning AI tools, if they inherit the structural inequalities of the past as a learning pattern, can reproduce gender biases and perpetuate situations of inequality, and even amplify them unnoticed by the person using them due to the appearance of technological neutrality.
This phenomenon, which is called “algorithmic discrimination”, is considered by European and Spanish regulations.
The regulatory system designed by Law 3/2007 and Royal Decrees 901/2020 and 902/2020 is widely known, which have introduced useful tools to achieve equality between men and women in companies: equality plans, remuneration registers, remuneration audits and job evaluation systems to identify those of equal value.
In the coming years, these regulations will be complemented by the transposition of the recent Directive (EU) 2023/970, aimed at strengthening pay transparency and the principle of equal pay, as it requires that the criteria that determine remuneration and promotion in companies be objective, neutral and understandable. In fact, this directive expressly requires that professional classification systems exclude all discrimination, including that resulting from the use of technological tools.
For its part, Directive (EU) 2019/1152 on transparent and predictable working conditions also requires a revisit of people management systems, and if these include algorithmic or AI tools, their objectivity cannot be presumed.
Finally, the Artificial Intelligence Regulation has classified as “high-risk systems” those used in the field of employment and people management about selection and hiring tools or decision-making on working conditions.
The confluence of both regulations —pay transparency and artificial intelligence— sets a new standard of diligence for organizations that use automated systems in talent management.
Some good practices to comply with this regulation to reduce the risk of algorithmic discrimination include:
- Auditing the AI systems and algorithms used in people management processes, to avoid potential biases derived from the learning of historical data.
- Integrating the gender perspective in the evaluation of positions and in the definition of professional careers and promotion criteria.
- Ensuring human oversight of automated decisions.
- Training teams in charge of talent management, in various facets, so that they are aware of the risks of AI and its impact on equality in the company.
In short, in the design and implementation of their human resources strategies, organizations must integrate regulatory compliance, technological innovation and gender equality.
