Artificial intelligence with perspective: transforming biases into opportunities for women and girls in technology

Artificial intelligence (AI) is increasingly present in our lives: from algorithms that filter content on social networks to recommendation systems of educational platforms, through health applications, job recruitment and transport. However, like any technology created by humans, AI is not neutral. One of the most important challenges facing this discipline today is that of algorithmic biases, which can amplify pre-existing inequalities if not properly addressed.

Paradoxically, these biases can also be a powerful lever of change: They make visible the existing gaps and offer us the possibility of using AI as a tool to create technological vocations in girls and women, a group historically underrepresented in science and technology.

What are AI biases?

AI biases arise when the data with which an algorithm is trained reflects stereotypes, inequalities or exclusions present in society. If an AI learns, for example, from employment data where historically men have been preferred for technical positions, it is likely to replicate that trend, discarding female candidates even if they are equally or better qualified.

A famous case occurred in 2018, when Amazon had to rule out an AI system that evaluated curricula because it systematically discriminated against women. The system had been trained with data from previous recruitments, dominated by male profiles, and penalised terms such as ‘women’s football team’ or the mention of women’s universities.

Why are there fewer women in technology?

According to Eurostat data (2023), only 19% of ICT specialists in the European Union are women. In Spain, the figure is around 17%. The causes are multiple and complex: gender stereotypes from an early age, lack of female role models in STEM (Science, Technology, Engineering and Mathematics), an uninclusive work environment and the absence of educational references that encourage girls to pursue technical careers.

This imbalance is not only unfair from an equity point of view, but it is also a practical problem: if women are not involved in the development of technologies such as AI, their needs, realities and perspectives will not be adequately represented.

Using biases as a mirror: an educational opportunity

Although AI biases are a problem, they also offer us a valuable pedagogical tool. Talking about bias is not just talking about technical errors: It's talking about how technology reflects our culture. This approach can be very useful in the classroom and in outreach activities to arouse the interest of girls and adolescents in technology.

For example, educational projects such as AI4ALL in the United States, or the Technovation Girls program in more than 100 countries, teach high school youth not only to program, but to think critically about how algorithms are built and to identify possible biases. In this way, girls not only learn to use technology, but to transform it from their own experiences.

In Europe, initiatives such as Digital Girls in Germany or the Ellas Lideran el Futuro programme in Spain are helping to reduce the digital gender gap from an intersectional and inclusive approach. And especially, in the Spanish case, highlights the work of the Inspiring Girls Foundation, which connects school-age girls with women referents of different professions, including those in the technological and scientific field. Through talks, workshops and mentoring, girls can visualize themselves in roles that previously seemed distant, and understand that they can also be programmers, engineers or experts in AI.

Showing how an algorithm can be unfair if you do not think about diversity awakens in many young people the desire to change the rules of the game.

AI needs more female voices

One of the main problems of today's AI systems is the lack of diversity in the equipment that designs them. According to a study by AI Now Institute (University of New York), 80% of AI researchers at major tech companies are men. And if we add that most are white and of similar socioeconomic contexts, bias is almost inevitable.

Including more women, and especially women from different backgrounds, is not just a matter of justice, but of improving the quality of technologies. Diverse teams better detect errors, design more inclusive products, and better understand the needs of the entire population.

A good example of this is the work of researcher Joy Buolamwini, founder of the Algorithmic Justice League project, who revealed how facial recognition systems failed more when identifying faces of black women. His research not only generated an academic impact, but forced large companies such as IBM and Microsoft to review their algorithms.

Five steps to transform AI into an ally of equality

1. Mainstream AI in education from an early stage, especially in gender-focused school contexts. Talking about ethics, biases and diversity in AI from primary or secondary can spark unexpected vocations.
2. Create visible referents: raise awareness of women leaders in AI and technology, from scientists like Fei-Fei Li to activists like Safiya Noble or local researchers working on issues relevant to girls.
3. Promote technological projects with social impact, where girls see that technology can serve to improve their communities, defend rights or solve problems that directly affect them.
4. Accompany with public policies that promote equity in access to technical and scientific training, specific scholarships and mentoring programs for girls and adolescents interested in technology.
5. Train teachers in gender and technology so that they are able to detect biases in their own practices and encourage their students with appropriate tools.

Artificial intelligence is neither good nor bad by itself. It is a powerful tool that can amplify inequalities or correct them, depending on who designs it and for what purpose. In this context, turning AI biases into an educational opportunity allows us not only to improve technological systems, but also to build a fairer and more inclusive future.

For this, it is essential that girls and women see themselves not as passive users of technology, but as active creators of a new digital paradigm. Because tomorrow's AI must speak with many voices, and among them, female voices are essential.