Addressing gender inequality in entrepreneurship requires looking beyond obvious barriers to examine how language itself shapes opportunities. By making these subtle patterns visible, this research offers a foundation for creating more inclusive entrepreneurial ecosystems that recognize and value diverse approaches to entrepreneurship.
In connection with Farnaz Aref's forthcoming defence of her PhD project “Words that shape opportunities: An AI-driven analysis of gender bias patterns in entrepreneurial discours", she has shared some of her main findings and how she believes they could benefit society.
Farnaz’ dissertation examines how gender bias operates through language in entrepreneurial contexts. Using advanced AI methods, she analyzed subtle linguistic patterns in entrepreneurial narratives and pitch competitions to reveal how bias manifests in ways often invisible to traditional research methods. By identifying these linguistic mechanisms, her research provides both a methodological framework for detecting bias at scale and evidence of how seemingly neutral language can create systemic barriers for women entrepreneurs.
A journey that began long before the official PhD scholarship
Farnaz’ PhD journey began long before she officially applied for a scholarship at DTU. When she was pregnant with her first daughter, she found herself researching gendered representations in media and the labor market, specifically in STEM fields, concerned about the stereotypes my daughter would encounter.
In her preface, she writes: "Today, entrepreneurship stands as a prestigious career path that many aspire to pursue. Yet paradoxically, the narrow portrayal of entrepreneurship in media and educational settings remains constrained by a dominant archetype—often characterized by aggressive growth strategies and profit maximization. These representations have dominated not only public discourse but also educational materials, potentially limiting our collective imagination about who can be an entrepreneur and what entrepreneurial success looks like”. She continues: ”This dissertation emerges against a backdrop of growing recognition that entrepreneurship should encompass broader definitions of success that include societal impact and ethical considerations. It also arises at a time of complex societal discussions about diversity and inclusion in business, where progress toward inclusivity is facing growing resistance”.
Women secure just 2% of venture capital funding
Despite decades of research on gender inequality in entrepreneurship, women still secure just 2% of venture capital funding. My research addresses a critical gap in understanding why these disparities persist by examining how language itself constructs and maintains gender bias in entrepreneurial settings. I developed an innovative methodological approach using Large Language Models to detect subtle and implicit biases in entrepreneurial discourse that traditional research methods often miss.
Three main findings
- Gender bias in entrepreneurship predominantly manifests through subtle linguistic mechanisms rather than explicit discrimination.
- In media narratives, female entrepreneurs are disproportionately featured in traditionally feminine industries, while entrepreneurship itself is framed through masculinized concepts.
- In pitching contexts, teams with higher proportions of female members systematically receive more prevention-focused questions about risks and potential losses, while male-dominated teams receive more promotion-focused questions about opportunities and growth.
Unveiling bias: Insights for scholars, investors, programme managers and policymakers
“For scholars, my work offers a methodological framework for analyzing bias at scale without sacrificing contextual sensitivity. For investors and program managers, it highlights specific questioning patterns to monitor in their evaluation practices. For entrepreneurs, understanding these linguistic mechanisms can help prepare for and navigate biased evaluations. For policymakers, my findings suggest that addressing structural biases in evaluation formats, rather than focusing solely on individual-level interventions, may more effectively create equitable entrepreneurial ecosystems”, Farnaz says.
A pathway for targeted interventions
By making subtle bias mechanisms visible, Farnaz Aref’s research opens pathways for targeted interventions that address root causes rather than just symptoms of gender inequality.
The methodological innovations demonstrated in her work show how AI can be leveraged as a tool for social science research, particularly in detecting patterns that might otherwise remain hidden. As entrepreneurial ecosystems increasingly recognize these linguistic mechanisms of bias, we can develop more inclusive evaluation practices and media representations that value diverse entrepreneurial approaches and ultimately create more equitable access to resources.
PhD defence
Farnaz Aref will defend her PhD project on 19 June 2025.
See location, summary and other details of her defence here
Reach out to Farnaz on LinkedIn
About research at DTU Entrepreneurship in general
We focus on evidence-based research to tackle a wide range of issues and provide opportunities for social and economic change.
Do you want to explore more about research at DTU Entrepreneurship
Reach out to Professor, Head of Research, Jason Li-Ying
Research areas