Teachers are increasingly using generative AI tools like ChatGPT-4 in K-12 classrooms worldwide. However, teacher education often lacks foundational knowledge in computer science (CS), particularly in AI, especially in the context of early education (K-4). To harness the potential of generative AI like ChatGPT-4 effectively, educators must be equipped with the necessary knowledge and strategies to integrate these tools into their teaching practice. This chapter presents an approach to differentiating tasks using ChatGPT-4. Building on prior research, several pedagogical criteria were selected to adapt existing tasks in order to address the diverse levels of students’ prior knowledge. This approach was piloted in two workshops with teachers and researchers from various backgrounds, including early education. Following a revision of the prompts, the method was tested again with pre-service teachers from a range of disciplines. Based on the insights gained from these trials, this chapter serves as a guide for early educators, offering a pedagogically grounded method to differentiate tasks for homework, classroom activities, and exams. By increasing the adaptability of their lessons, educators can better address students with varying levels of prior knowledge. The limitations of this approach are discussed at the conclusion of the chapter.