Illusions of Understanding: The Epistemic Risks of Artificial Intelligence in Software Engineering Research

This study investigates the growing integration of artificial intelligence (AI) in scientific research, focusing on its application within software engineering. As AI becomes a more prominent tool for enhancing productivity and objectivity, concerns emerge regarding its potential to create illusions of understanding. This research examines how AI-driven monocultures in scientific practices may narrow the scope of inquiry and marginalize alternative perspectives, ultimately limiting innovation. Through a critical analysis of AI’s roles—Oracle, Surrogate, Quant, and Arbiter—this paper explores the balance between AI’s promises and epistemic risks, concluding with recommendations for fostering diverse and responsible AI use in software engineering research.