Although the teaching profession is largely chosen by women worldwide, the insufficient representation of women in leadership and administrative positions in education constitutes a striking paradox. Across OECD countries, while nearly 69% of teachers are women, only 45% of school principals are female, revealing a sharp decline in leadership representation (OECD, 2017). In Turkey, the percentage of female rectors was 5.3% in 2007 and increased to 9% in 2023. Meanwhile, the ratio of female vice rectors was 4% and female deans was 17% in 2023 (Turkish Higher Education Management Information System, 2023). According to the 2023 report by the Ministry of Family and Social Services, only 3 out of 81 Provincial Directors of National Education are women, representing 3.7% of these positions.

Butler (1990) argues in her book, Gender Trouble that gender is “performative”; In other words, gender is not an innate characteristic, but a process constantly reproduced through culturally determined actions. People perform these performances unconsciously, and their identities are shaped by these repetitive actions. Butler emphasizes that gender is constructed through linguistic and cultural discourses, and that no meaning related to the body predates social structure. Therefore, she denies that the biological body or the fetus exists with "natural" meanings. In short, according to Butler, gender is not something a person "is," but something they "do."

The rationale for this study is to understand how artificial intelligence models create language and discourse about female school leaders. Generative AI systems reproduce patterns they learn from data sources. Therefore, their responses to female leadership may be influenced by social perceptions, cultural norms, and common expressions. Examining AI's representations in this field can contribute to both how gender perceptions are constructed in digital environments and the forms in which potential biases emerge. The purpose of this study is to examine how female school leaders are represented in generative AI algorithms on a global scale. In this context, the research aims to analyze how different generative AI models generate discourses about female school leadership and the gender norms these discourses implicitly or explicitly reproduce or transform.

Research problems:

According to artificial intelligence algorithms, how are women school leaders defined, and what is the representation of women leaders in educational institutions globally?

According to artificial intelligence algorithms, what are the main cultural, structural, and institutional factors contributing to the underrepresentation of women in school leadership roles in certain countries?

According to artificial intelligence algorithms, what are the effects of the low representation of women school leaders on education systems, school performance, and various school-related variables?

According to artificial intelligence algorithms, what positive outcomes might an increase in women school leaders generate for educational organizations, what leadership style trends are observed, and what differences exist between women and men leaders?

According to artificial intelligence algorithms, how is the role of women school leaders expected to evolve in the future, what policy and structural transformations are needed to strengthen women’s leadership, and how might AI reshape the representation of women’s leadership in education in the digital age?

Qualitative research methods will be preferred in the research. Qualitative research is a type of inquiry that employs methods such as observation, interviews, and document analysis to address a research problem (Seale, 1999). In data analysis, the critical discourse analysis method will be used. Critical discourse analysis is an approach that examines the effects of language on social structure and power relations. This approach analyzes how concepts such as ideology, power, and identity are shaped and reproduced through discourse and texts. It also supports social change by revealing injustices and social inequalities (Van Dijk, 2001). In this study, questions will be posed to the ChatGPT, DeepSeek, Gemini, Mistral, Claude, Kumru, Qwen, and Alice AI models. These AI models were selected due to their cultural diversity, technical differences, and dataset diversity. This diversity can provide the opportunity to examine how discourses regarding representations of women in educational leadership differ, what ideological influences they carry, what prejudices they reproduce, and which models produce more egalitarian discourse. In this context, the questions to be posed to artificial intelligence models are as follows: How do you define women school leaders? What is the representation of women leaders in educational organizations worldwide? What are the reasons for the underrepresentation of women in school leadership roles in some countries? What are the potential impacts of the low number of women school leaders on education systems in some countries? What positive outcomes could an increase in women school leaders create in these systems? What might be the salient professional and personal characteristics of women school leaders? What are the common trends regarding the leadership styles of women school leaders? What differences do you observe between women and men school leaders? What might be the impact of women school leaders on school variables? How do you assess the future evolution of the role of women school leaders in education? What policies or structural transformations might be necessary to strengthen women school leadership? What are your thoughts on AI transforming representations of women's leadership in education in the digital age? According to Ravitch and Carl (2019), validity and reliability procedures will be conducted in the study. For credibility, expert opinions will be taken. Confirmability, findings based on algorithms' statements will be emphasized. Transferability refers to the applicability of the findings to similar contexts.

AI models are expected to mostly portray female school leaders as individuals with strong communication skills, empathy, openness to collaboration, and strong capacity for community building. While these representations are positive, they can sometimes reproduce normative gender-based roles. Models may tend to indicate that the proportion of female school leaders worldwide is lower than that of men. They are also expected to explain this situation with structural factors such as cultural norms, gender roles, patriarchal institutional structures, and career-family incompatibility. AI models are expected to highlight the negative consequences of a low number of female leaders, such as a lack of diversity and inclusiveness, a weakening of school climate and communication quality, and uniformity in decision-making processes. Models focusing on the positive effects of increased female leadership are expected to highlight the following positive outcomes: a more democratic and inclusive school culture, increased teacher motivation, enhanced student well-being and school engagement, and a role model role in gender equality. Some discourses within the models may create the risk of unwittingly framing female leadership as "emotional" and male leadership as "rational." This may indicate the presence of implicit gender stereotypes in algorithmic discourses. AI models are expected to generate discourses generally optimistic about the strengthening of female leadership in the future, such as policies improving and public awareness rising. The following recommendations can be developed within the scope of the study: More in-depth discourse analyses can be conducted to reveal implicit biases in representations of female leadership in generative AI models. Gender theories can be rethought through digital discourses, and conceptual frameworks can be developed to understand how AI reimagines gender. Mentoring and leadership development programs can be strengthened to increase the visibility of female school leaders. Regulations focused on gender equality can be developed for leadership appointments in education.

Butler, J. (1990). Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. OECD. (2017). Education at a Glance 2017: OECD Indicators. OECD Publishing. Ravitch, S. M., & Carl, N. M. (2019). Qualitative research: Bridging the conceptual, theoretical, and methodological. Sage Publications. Seale, C. (1999). Quality in qualitative research. Qualitative Inquiry, 5(4), 465-478. Turkish Higher Education Management Information System (2023). https://istatistik.yok.gov.tr Van Dijk, T. A. (2001). Multidisciplinary CDA: A plea for diversity. In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis (pp. 95-120). http://dx.doi.org/10.4135/9780857028020.d7