The Wade Test: Generative AI and a CEO Bot

CESifo

Abstract

This study investigates whether Gen-AI can mimic a human CEO and whether employees display aversion to Gen-AI communication. We presenta framework of Gen-AI aversion that distinguishes between a strong form (unrelated to Gen-AI performance) and a weak form (related to Gen-AI’s performance), and potential Gen-AI failures. In a field experiment with a firm, we extend the idea of a Turing test (i.e., a computer mimicking a person), tothe idea of AI mimicking a specific person, namely the CEO. We call this the “Wade test” and assess if employees can distinguish between communication from their CEO and communication generated by an AI trained on the CEO’s prior communication. We find that responses are correctly identified 59% of the time, slightly better than random chance. When employees believe a response is AI generated, regardless of its actual source, they perceive it as less helpful. A second study with a general audience, using public statements from CEOs and from an AI intended to mimic those CEOs, finds that AI-labeled responses (irrespective of their actual source) are rated as less helpful. Taken together, our findings provide evidence of strong algorithm aversion to the use of Gen-AI in impersonating communication.

Authors 3

1. Prithwiraj Choudhury
2. Bart S. Vanneste

   University College London

   Affiliation as printed

   UCL School of Management University College London

3. Amirhossein Zohrehvand Aachen Faculty of Science, Science Based Business Leiden University

   Leiden University

   Affiliation as printed

   Faculty of Science , Science Based Business Leiden University