The purpose of this viewpoint is to address the often unclear definition of artificial intelligence (AI) in research, arguing that researchers must clearly specify AI system types and novel characteristics when examining human perception to avoid repeating known results with new labels.
The analysis identifies key distinguishing dimensions between AI and traditional software: probabilistic decision-making versus rule-based logic, task adaptability, conversational collaboration modes, emotional relationship potential through natural language interaction, environmental integration and co-evolutionary learning dynamics.
Generative AI transforms human–machine interaction through continuous availability and perceived partnership, creating fundamentally different user experiences compared to conventional software systems.
Researchers should explicitly describe AI embodiments to study participants and verify comprehension with control questions, thereby enabling accurate assessment of perceptual differences and meaningful theoretical contributions beyond merely relabelling existing software research.
A precise AI specification enables practitioners to understand genuine technological advantages, while researchers can develop rigour results that specifically address AI’s unique characteristics.
The framework provides specific criteria for distinguishing AI novelty from traditional software in perception studies, addressing widespread conceptual ambiguity that undermines research validity.
