Challenging the assumption that more knowledge leads to greater willingness to adopt AI, this study investigates how people can be biased in adoption decisions, testing the impact of Dunning-Kruger effect, confirmation bias and framing effect on the intention to use AI and if this relationship is moderated by different informational sources. Most AI adoption theories assume perfect rationality; humans, however, are not perfectly rational: knowledge can be biased, influencing behavioral intention. We conducted a survey-based experiment with a 2×2 between-subjects design, with undergraduate students enrolled in an AI course. We found no significant relationship between objective AI knowledge and intention to use AI. In contrast, AI self-efficacy and prior beliefs emerged as significant predictors of intention to use. We also found that formal and informal information sources moderate the relationship between self-efficacy and intention to use AI, strengthening it in the case of formal sources and weakening it in the case of informal sources. These findings extend research on AI adoption and offer practical insights for managers seeking to reduce resistance to AI use.
Giorda, F., Iacopino, V., Gabbriellini, S., Rajola, F., Investigating the impact of cognitive biases on intention to use AI and the moderating role of informational sources: evidence from a survey-based experiment, in AMCIS 2026 Proceedings, (Reno, Nevada, 20-22 August 2026), AMCIS American Conference on Information Systems, Reno, Nevada 2026: 1-10 [https://hdl.handle.net/10807/351148]
Investigating the impact of cognitive biases on intention to use AI and the moderating role of informational sources: evidence from a survey-based experiment
Giorda, Federica
Primo
Conceptualization
;Iacopino, ValentinaSecondo
Conceptualization
;Gabbriellini, SimonePenultimo
Formal Analysis
;Rajola, FedericoUltimo
Supervision
2026
Abstract
Challenging the assumption that more knowledge leads to greater willingness to adopt AI, this study investigates how people can be biased in adoption decisions, testing the impact of Dunning-Kruger effect, confirmation bias and framing effect on the intention to use AI and if this relationship is moderated by different informational sources. Most AI adoption theories assume perfect rationality; humans, however, are not perfectly rational: knowledge can be biased, influencing behavioral intention. We conducted a survey-based experiment with a 2×2 between-subjects design, with undergraduate students enrolled in an AI course. We found no significant relationship between objective AI knowledge and intention to use AI. In contrast, AI self-efficacy and prior beliefs emerged as significant predictors of intention to use. We also found that formal and informal information sources moderate the relationship between self-efficacy and intention to use AI, strengthening it in the case of formal sources and weakening it in the case of informal sources. These findings extend research on AI adoption and offer practical insights for managers seeking to reduce resistance to AI use.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



