Developing the CHAT Framework for Ethical Instructor-AI Integration: A Mixed Methods Study in Trinidad and Tobago

Purpose: This study introduces the CHAT (ChatGPT, Holistic, Adaptive, Teaching) Framework, a contextually grounded model that integrates UTAUT, TPACK, and UNESCO’s AI in Education Policy to support ethical and effective instructor-AI collaboration. It investigates how higher education instructors in Trinidad and Tobago perceive, adopt, and integrate generative AI tools, particularly ChatGPT, within their teaching and administrative practices. Method: An explanatory sequential mixed-methods design was employed. A quantitative survey of 101 instructors was followed by 11 semi-structured interviews and one focus group (n = 6) to explain statistical results. Quantitative data were analysed using descriptive statistics, Spearman’s rho, and one-way ANOVA. Qualitative data were thematically analysed to interpret and expand quantitative findings. Findings: Results indicated moderate adoption of ChatGPT (M = 3.47, SD = 0.81), with effort expectancy (ρ = .62, p < .01) and facilitating conditions (ρ = .57, p < .01) showing the strongest positive correlations with adoption. No significant differences emerged across gender, age, or education level (p > .05). Qualitative findings revealed enthusiasm for AI’s instructional efficiency but raised ethical concerns about plagiarism, bias, and data privacy. Value: This study contributes an empirically grounded, human-centred framework that explains how adoption conditions, pedagogical practice, and ethical mediation interact to shape instructor-AI integration in Caribbean higher education. Practical Implications: The CHAT Framework offers policymakers, administrators, and educators a human-centred model to guide responsible professional development initiatives and institutional strategies for equitable and sustainable AI use.