Modelling the Antecedents of Perceived Trust in AI-Based Records and Information Management Systems

This study explores the antecedents of trust and its impact on the intention to use artificial intelligence (AI)-based records and information management systems (RIMS), applying the trust-based extended valence framework (TEVF) to an African context, where research on AI trust in records and information management remains limited. Using a mixed methods approach, snowball and purposive sampling techniques were used to gather data online from 890 participants. Correlation analysis and ANOVA were conducted to examine the TEVF dimensions, while thematic analysis was applied to qualitative data. The findings revealed that perceived trust was positively correlated with behavioural use intention and perceived benefits. However, perceived risks was not significantly correlated with either behavioural use intention or perceived benefits. The relationship between perceived trust and perceived risks was found to be weakly negative. Perceived trust and behavioural use intention exhibited significant group differences based on geographic location and work experience. Transparency, understanding basic functionality, human oversight and control, ease of use, explainability, and regulatory framework were identified as key trust enablers, while cybersecurity concerns, ethical risks, resources constraints, and skills constraints emerged as major barriers to behavioural use intention. The study posits that sustainable AI integration into records and information management hinges on transparent governance, targeted training, and strategic advocacy, ensuring efficiency while mitigating adoption barriers.