[Purpose/significance] With the rise of generative AI, its role as a social actor is reshaping the process of in⁃formation acquisition. Against the backdrop of the uncertainty arising from human to human proxy search, exploring the influence of uncertainty on human to machine proxy search behavior from the perspectives of positive and negative emo⁃tions is of great significance for the further development of this technology. [Method/process] Based on the motivation⁃al information management theory, a research model was constructed, and a total of 13 research hypotheses were pro⁃posed. The data were collected using the questionnaire survey method and the critical incident technique. The relevant model calculations were carried out through the partial least squares structural equation modeling. [Result/conclu⁃sion] This study first reveals the correlation between the uncertainty differences, the low matching between information and tasks, and individual emotions. That is, an increase in uncertainty differences enhances individuals' negative emo⁃
tions and, correspondingly, reduces positive emotions. Under the influence of negative emotions, users lower their eval⁃uation of expected outcomes and effectiveness, thereby suppressing human to machine proxy search behavior. Although the low matching between information and tasks has no significant impact on the emotional state, it positively promotes human to machine proxy search behavior. Emotions have no significant impact on the results of human to machine proxy search behavior.