The COVID-19 pandemic has transformed teaching and learning in ESL education, particularly through the integration of artificial intelligence (AI) technologies into current English Language Teaching (ELT) practices. However, to ensure the seamless integration of these technologies into ELT practices, a proper framework needs to be developed and validated for future reference by ESL educators. Hence, this study aims to develop and validate the proposed constructs for the framework, using the Content Validity Index (CVI) method with assistance from six experts. Eleven constructs were identified from previous studies, with seven derived from the AI-TPACK framework and four from the Framework for Teaching. These constructs are categorised into three domains: Technological Proficiency, Pedagogical Compatibility, and Social Awareness. The data were analysed using the CVI method. The results show that the S-CVI/ Ave value was 0.91, meeting the satisfactory CVI threshold for six experts (>0.83). However, the S-CVI/UA value was 0.73, which does not meet the acceptable CVI value (>0.83). As a result, 10 items with an S-CVI/UA less than 0.83 were removed. Additionally, two more items were excluded due to their constructs having insufficient valid items, with only one item remaining in each construct. Following these amendments, both the S-CVI/Ave and S-CVI/ UA values improved to 0.99 and 0.91, respectively. The study proves that the content validity process is crucial to ensuring a reliable and valid instrument for developing a framework for integrating AI into the ELT practices.