The double-edged sword of AI dependency: how cognitive offloading and skill atrophy foster digital complacency

Purpose The current study examines the cognitive implications of an addictive use of generative artificial intelligence (AI) tools, and it theorizes that AI dependency triggers a cascade of mediation involving cognitive offloading and skill atrophy that results in digital complacency. Using the tenet of Distributed Atrophy Model, the study develops and tests an integrated model. Design/methodology/approach The data were collected through an online cross-sectional survey from 303 professionals (mostly aged 25–45) across technology-intensive and other AI-adopting knowledge work sectors (Information Technology, Higher Education, Finance and Banking, and Consulting) who use generative AI on a regular basis in their work-related activities. The proposed model was tested by using structural equation modeling (SEM-AMOS) that entailed confirmatory factor analysis of measurement validity and bootstrapped path analysis (5,000 resamples) to assess direct and mediated effects. Findings The results revealed that AI dependency positively predicted cognitive offloading, which had a strong impact on skill atrophy and skill atrophy was the strongest predictor of digital complacency. The direct impact of AI dependency on digital complacency was significant, which was serially mediated by cognitive offloading and skill atrophy. The model accounted for 36% of the variance in digital complacency. Originality/value Since the emergence of AI, most research has mainly focused on the positive implications of technology. But there are new indications that AI has the capacity to produce negative outcomes as well. Furthermore, the available literature is still fragmented, and various paths have not been thoroughly studied. To fill this gap, the current study discusses the negative consequences of AI dependency by developing and testing an integrated model to offer a more detailed insight into its cognitive and behavioral impacts.