Abstract
<p>The Ethical Epistemic Collaborative Output Dialogue (EECOD) is a structured classroom protocol designed to support AI fluency and make student thinking visible throughout the learning process. Developed from Further Education and Training (FET) practice, EECOD guides students through a standardised interaction with large language models (LLMs) that is transparent, assessable, and teacher-verifiable. Students explain the task in their own words, receive one constrained LLM suggestion at a time, and accept or reject each with a stated reason. This visible chain of reasoning is captured in a Protocol Score Card that evidences both student engagement and teacher intervention without adding to workload. The framework is mapped against Irish National Framework of Qualifications (NFQ) and Quality and Qualifications Ireland (QQI) standards using Bloom’s Revised Taxonomy. Worked examples and a proposed assessment rubric are provided. The paper argues that EECOD represents a replicable, subject-transferable model for process assessment in an era when AI use in education is unavoidable but largely unmediated.Keywords: AI fluency; visible thinking; assessment for learning; Teacher in the Loop; large language models; academic integrity</p>