This study explores EFL educators’ perspectives and practices of using ChatGPT for materials development in higher education. The research was conducted with 60 faculty members from Sultan Qaboos University using a mixed-methods design. Guided by Rogers’ Diffusion of Innovations theory, the study examined relative advantage, compatibility, complexity, trialability, and observability of ChatGPT’s affordances and inadequacies in materials development. Data were collected using a structured questionnaire with Likert-scale items and open-ended questions, designed to capture both quantitative trends and qualitative insights. Descriptive statistics were used to analyze survey responses, while thematic coding was applied to open-text data, ensuring alignment with the study’s theoretical framework. Results show that teachers view ChatGPT as efficient, flexible, and supportive of curriculum alignment. They valued its ability to save time and generate adaptable content. However, concerns about accuracy, cultural relevance, and policy clarity were common. Faculty reported revising outputs to fit learners’ needs and stressed the importance of institutional support and training. The findings highlight opportunities for responsible integration of ChatGPT in ELT, while also noting risks of over-reliance. This study contributes to understanding how AI can reshape materials development and offers guidance for teacher training and policymaking in emerging educational contexts.