GenAI and Writing Age – Can we use it in detection of use?

This paper is the first step in answering the question can the authenticity of student work be confirmed, or challenged through assessing the writing age of submitted work. There have been numerous papers published on the subject of authenticity of student work with none offering very much hope for a robust solution. The work undertaken in this project has not been undertaken with the intention of trying to create a robust detection method but rather to explore the question of writing age which itself is a well established text analysis field. The project set out to ask the following questions: "What age does GenAI write at?" and "If a GenAI tool is prompted to write for a given age band how accurately does it achieve this?" The project used three commonly available GenAI tools and each was prompted to write a short assignment using the same prompt. The prompt was then modified to specify one of three target writing ages – University entry level (A level in the UK), Undergraduate and Postgraduate. The results suggests reading age analysis is an area worth exploring in detecting GenAI generated content in submitted assignments but may still be dependent on having a sample of authentic student writing as a comparator and a good understanding of the acceptability of AI as a proofreading and/or grammar improving tool. The study is limited in extensiveness the findings are however not promising.