On February 27, 2026, a regular research seminar of the International Laboratory of Intangible-driven Economy (IDLab) was held. Valeriia Fedorova, Research Assistant at IDLab, presented a talk titled “Use of Generative AI in Project Descriptions and Funding Outcomes in Crowdfunding.”
The study was co-authored by Petr Parshakov, Head of the Laboratory, and Marina Zavertiaeva, Senior Research Fellow at IDLab.
.jpg "IDLab Seminar: The Impact of AI-Generated Project Descriptions on Crowdfunding Campaign Outcomes")
The study focuses on an increasingly relevant question: Does the use of generative artificial intelligence (AI) in writing project descriptions affect the outcomes of crowdfunding campaigns? In other words, do audiences perceive “human-written” and AI-generated texts differently - and does this difference impact fundraising success?
The empirical analysis is based on data from the Kickstarter platform. In the first stage, the researchers trained a specialized machine learning model capable of detecting whether a project description was generated with the help of generative AI. They employed the RoBERTa architecture, training the model on projects published before 2021 alongside AI-rephrased versions of those same texts. This approach enabled them to estimate the likelihood of AI use in campaigns launched in 2023–2024.
In the second stage, the authors examined how detected AI usage correlates with campaign financial performance. Regression analysis revealed a negative relationship: projects with a high probability of AI-generated descriptions were approximately 16% less likely to reach their funding goal and, on average, raised 66% less money.However, the effect was not uniform across all campaigns. Quantile regression showed that the negative impact is significantly weaker for the most successful projects. In other words, large-scale campaigns appear less sensitive to whether the text was written by a human or generated by an algorithm.
The seminar concluded with a particularly lively discussion. Participants proposed several directions for future work—most notably, a more detailed investigation into whether the effect varies by project type (e.g., technological, creative, social, etc.). These suggestions could serve as a foundation for deepening the analysis and expanding the empirical scope of the study.