Soutenance de thèse de BASHARAT Iqra

Résumé de la thèse

Artificial intelligence (AI) is increasingly transforming entrepreneurial ecosystems by enabling startups to innovate and grow quickly. At the same time, growing ethical concerns related to transparency, accountability, and responsible AI have increased the importance of ethical legitimacy for AI-based companies. Despite growing academic interest, the existing literature remains fragmented and offers a limited understanding of how AI startups communicate their ethical responsibility and how stakeholders interpret ethical signals in contexts of uncertainty. This thesis addresses this gap by examining how AI-related ethical signals influence stakeholder evaluation, investment decisions, and the success of crowdfunding in AI-driven entrepreneurship.
Relying mainly on signal theory, this thesis adopts a multi-method methodological approach composed of three interconnected trials. The first essay conducts a systematic literature review to synthesize research on AI-based startups, ethical practices, ethical certifications, and investor decisions. The second trial mobilizes three experimental studies to examine how third-party ethical certification and voluntary disclosure of ethical information influence willingness to invest in AI startups in the healthcare sector. The results show that third-party ethical certification increases willingness to invest by reducing perceived ethical risk, while voluntary disclosure weakens the credibility of the signal and reduces this willingness to invest. The third essay analyzes AI-related ethical communication through machine learning (ML)-assisted textual analysis of 2,008 AI-related Kickstarter campaigns. The results reveal an inverted U-shaped relationship between the intensity of AI-related ethical language and crowdfunding success: moderate ethical communication improves campaign performance, while excessive ethical reporting leads to stakeholder skepticism. External credibility signals mitigate these negative effects.
This thesis contributes to signal theory by demonstrating that ethical signals are heterogeneous, nonlinear and dependent on signal credibility and external validation. She also contributes to the literature on AI-based entrepreneurship by positing ethical legitimacy as a strategic entrepreneurial resource, influencing stakeholder trust and resource acquisition. Methodologically, the thesis integrates a systematic review of literature, experiments and machine learning-assisted archival analysis in order to examine ethical signals at different levels of entrepreneurial activity. The thesis thus makes theoretical and practical contributions to the understanding of how AI-based startups manage ethical legitimacy challenges in increasingly monitored technological environments.