The Role of Review Sidedness in AI vs. Human-Generated Summaries: An ELM Perspective on Information Processing Effort

This study investigates how the source (AI vs. Human) and sidedness (One-sided vs. Two-sided) of review summaries affect consumer information processing effort (IPE) in the context of online hotel booking. Drawing upon the Elaboration Likelihood Model (ELM), two experimental studies were conducted to measure IPE using two indicators: comprehension score and reading time. Study 1 reveals that review sidedness, rather than review source, significantly influences cognitive effort, with one-sided AI-generated reviews requiring the highest processing effort. Interestingly, for two-sided reviews, human-generated summaries demanded higher IPE than AI-generated ones. Study 2 finds that participants spent more time reading one-sided reviews than two-sided ones. These findings support the moderating role of review- sidedness in the relationship between review source and IPE. Additionally, the results highlight that perceived insufficiency and ambiguity in review content trigger deeper cognitive processing. Notably, confirmation bias and perceived human superiority also contribute to shaping consumers’ cognitive effort with AI versus human-generated content. This study contributes to existing literature by introducing review- sidedness as a moderating variable in AI-human content comparisons and offers practical implications for designing persuasive and trustworthy review systems in AI-assisted platforms.