Not all AI is created equal: A meta-analysis revealing drivers of AI resistance across markets, methods, and time

Abstract

While artificial intelligence (AI) is used by billions of consumers daily through tools like ChatGPT, prior research often documents that consumers are resistant to it. The current research proposes that such resistance is strongly context-dependent, rapidly evolving, and often an artifact of how researchers study it. We provide a comprehensive synthesis of consumer responses to AI by analyzing 440 effect sizes from 76,142 unique participants across two decades of experimental research. Our meta-analysis reveals three key insights about consumer aversion towards AI (average Cohen’s d = −0.21). First, consumer responses vary systematically by AI label and domain, with the most negative responses to embodied forms of AI (e.g., robots) compared to AI assistants or mere algorithms. We also identify substantial domain differences in areas such as transportation and public safety, which trigger more negative responses compared to areas where AI improves productivity and performance, such as in business and management. Second, we document a temporal evolution towards increasingly less negative responses, particularly for cognitive consumer responses (e.g., performance or competence judgements), with aversion approaching a null-effect in most recent years. Third, we demonstrate overall shrinking effect sizes with greater ecological validity. This work advances our understanding of when and why consumers resist AI and provides directions for future research on consumer-AI interactions.