Who trusts the bot more as a collaborator?

Generative AI (GenAI) offers transformative potential in education, yet the real-world adoption of (Gen)AI in education remains slow and uneven. This study aims to validate a new instrument measuring educators’ perceived cognitive and socio-affective trust in GenAI as a collaborator and to investigate if this trust varies across teaching levels, experience and academic qualifications. Inferential analyses such as One-Way ANOVA, nested ANOVA and General Linear Modelling (GLM) were conducted on self-reported data from 212 educators in Singapore. The findings revealed that educators with doctoral degrees in professional/adult education demonstrated the highest cognitive and socio-affective trust, while educators at the primary education level reported the lowest. These findings challenge previous research that reported no significant demographic or professional differences in trust in AI, highlighting the need for designing GenAI and its adoption strategies that are tailored to educators’ professional context and training. By identifying educators’ trust as a key factor in GenAI adoption, this study advances a human-centred perspective on trust in GenAI, particularly within the context of human-AI collaboration in teaching practice.