Recently, the collaborative paper AI helps until you say it does by Professor Niu Xiaofei from the School of Economics and the Center for Neuroeconomics has been officially published in The Scandinavian Journal of Economics, an international economics journal. The co-authors are Wang Yue, a PhD candidate, Professor Li Jianbiao from the School of Economics and the Center for Neuroeconomics, Shandong University, and Associate Professor Cao Qian from the School of Business Administration, Nanjing University of Finance and Economics.
With the rapid development of generative AI technologies, large language models such as ChatGPT are widely adopted in content creation, commercial services, public welfare communication and other fields. When the public is informed that content is generated or edited by AI, does such transparency disclosure affect people’s behavioral decisions? To address this question, the research team conducted a series of online experiments, recruiting approximately 4,300 participants from China and the United States. In the experiments, researchers designed different types of charitable fundraising letters, including texts fully generated by GPT, GPT-generated texts revised by humans, fully human-written texts, and texts polished with GPT assistance. They further manipulated whether participants were informed of AI involvement.
The results show that AI-generated fundraising letters can significantly boost participants’ donation behavior, with effects comparable to those of human-written letters. Even after AI-generated texts are manually edited, they retain positive effects. However, once participants are told that the fundraising letters involve AI generation or AI editing, this positive effect is markedly weakened or even disappears. Further analysis reveals that undisclosed AI-authored fundraising letters are more likely to evoke positive emotions and empathy among participants. In contrast, AI disclosure may reduce public perceptions of textual authenticity, thereby undermining its emotional impact.
This study identifies an important tension in the application of generative AI: AI technologies can improve the efficiency of information production and communication effectiveness, yet transparent disclosure of AI involvement may alter public judgments of content authenticity and consequently shape behavioral outcomes.
Niu Xiaofei holds a Doctor of Management degree. He is a professor (promoted via special appointment) and doctoral supervisor at the School of Economics and the Center for Neuroeconomics, Shandong University. He is a Young Taishan Scholar of Shandong Province (selected via special appointment) and a Qilu Young Scholar of Shandong University. His research covers neuroeconomics, behavioral and experimental economics, and behavioral corporate governance. As the first or corresponding author, his works have appeared in top international economics and management journals including Management Science (2 papers), Experimental Economics, and Journal of Economic Behavior & Organization (3 papers), as well as top multidisciplinary journals such as Proceedings of the National Academy of Sciences. He presides over one Youth Project of the National Natural Science Foundation of China and one sub-project of the Major Project of the National Social Science Fund of China.
Link to the original article: https://onlinelibrary.wiley.com/doi/10.1111/sjoe.70044
