EU struggles against gender prehistorization in artificial intelligence!

EU struggles against gender prehistorization in artificial intelligence!

Polen, Land - The European Union plans comprehensive measures to promote gender equality in artificial intelligence (AI). A draft that was presented by the Polish Minister of Equal Opportunities Katarzyna Kotula specifically names the need to act against gender-specific bias in AI systems. This is considered a crucial step, since there is concern that AI not only takes over existing gender stereotypes, but could also strengthen them. Stereotypes, such as the view that women act mainly in nursing professions and men are considered enforcement strengthers, are of central importance. The basis of these bias is often stereotypical training data that are used in the AI algorithms, which can lead to dangerous distortions. The situation in the healthcare system is particularly worrying, where data collected for male patients is often used for the training of AI systems, which disadvantaged women.

In order to meet these challenges, the EU Council asks Member States to expand their equality structures and create new authorities. It is planned that existing positions are strengthened, which includes a clear responsibility and more staff and budget. The goal is to establish AI as a tool for equality instead of acting as a vehicle for discrimination. In addition, a focus is on combating online misogynia, and it is recommended to examine "masculinist networks". Men and boys are to be actively involved in equality initiatives in order to promote sustainable change. This is particularly relevant because recent studies show that trust in AI varies between the sexes; Only 32% of women trust AI systems, while the proportion of men is 47%.

fight against discrimination

In addition, the draft addresses discrimination that women and girls experience in the digital world. Online violence in particular meets women in power positions and young women from marginalized groups, including politicians and journalists. The previous developments in the field of AI show that these technologies are often trained with distorted data that cement existing gender roles. For example, AI translations often bring stereotypical assignments to professions-while a nurse is typically identified as female, a doctor is almost always presented as male. This tendency increases the existing gender -specific clichés and leads to systematic distortions in thinking about gender roles.

The EU Council has therefore created legal foundations that are intended to integrate a comprehensive “gender perspective” into all political measures and laws. This also affects the AI Act and the Digital Services Act. This wants to ensure that gender justice is used not only in theory, but also in practice. In addition, a new guideline comes into force in December 2024, which provides for a quota for women for supervisory board positions in large listed companies. This promotes the gender-friendly line-up by at least 40% of non-managing directors must be women, or 33% of all supervisory bodies.

The way to the future

The report of the European Institute for Gender Course (Eige) This year emphasizes the need to reduce gender -specific stereotypes, which are underrepresenting women in particular in future industries such as AI and renewable energies. To make this possible, a strategy to close the gap between the views of young women and men must be developed for gender equality. The integration of a gender perspective can help reduce discrepancies in use and the skills in dealing with AI systems. At the moment, 40% of women do not use AI, while it is only 31% for men.

The discussion about gender equality in the AI is therefore not only a question of fairness, but also one that reflects and influences the social structures as a whole. Trade unions and organizations such as ver.di see the responsibility to drive fair conditions in the use of AI in the world of work. In order to identify and dismantle discrimination and prejudices in these systems, suggestions such as data filtering, human feedback and test results are recommended on stereotypes before new AI models are published.

The EU has taken the opportunity to make digital transformation gender -friendly and to actively reduce the inequalities prevailing in digital space. It remains to be hoped that these measures will not only have a sectoral, but also across generations.