Scaffolding reading can be challenging, especially for busy teachers. One major challenge is providing students with enough encounters with new vocabulary across varied and meaningful contexts in order for learning to occur. One approach that has become popular in Comprehensible Input–based teaching approaches is embedded readings developed by practitioner Laurie Clarcq. These are a series of leveled texts that recycle the same target vocabulary in meaningful ways while gradually increasing complexity and difficulty. When creating embedded readings, AI can be a powerful tool to support learners in acquiring language and save teachers time. In this presentation, we will show how AI can be used to quickly create embedded readings, how teachers can guide students to use AI appropriately for this purpose, and how these tools can support language learning. We will also compare embedded reading to other reading passages created by AI, emphasizing tradeoffs. Finally, we will end by discussing possible challenges, such as AI hallucinations and maintaining effective vocabulary control and repetition.