Research collaborator: AI and the cognitive work of writing
Education
I’m building Write with Milo, a writing support tool for students in grades 3–8 who get stuck when trying to turn their thoughts and ideas into writing, particularly students with dyslexia and language-based learning difficulties.
The idea began with something I saw as a teacher many years ago: a tool that recorded a student’s writing process in real time. I could see where a student stopped, restarted, changed a word or rewrote a sentence. It made visible something teachers rarely get to see unless we are sitting right beside a student—the process of writing itself.
For struggling writers, a finished piece can hide much of what happened along the way. A student may lose a thought mid-sentence, struggle to retrieve a word, repeatedly restart, or choose a simpler word because it is easier to spell. They may have ideas they can explain aloud but struggle to hold onto, organize and translate those ideas into written language.
We already have assistive technology that helps students access writing, tools that help improve a draft, and now generative AI that can produce the writing itself. I’m interested in a different question: Can we use AI to support the cognitive work of writing without doing that cognitive work for the student?
Writing is one way young people make meaning of themselves and the world around them. The words they eventually put on the page are only part of that process. Much of the important work is invisible: generating ideas, making connections, organizing thoughts, retrieving language, deciding what matters and figuring out how to communicate it.
As AI becomes increasingly capable of doing those things for us, I think there are important questions about where scaffolding ends and cognitive substitution begins. Which parts of the struggle involved in writing are barriers we should help students overcome? Which parts are productive and important to learning and development? And what might be the unintended consequences if AI routinely does that thinking for developing writers?
I’m looking for a researcher who is curious about these questions too—particularly someone working in writing development, learning sciences, cognition, learning disabilities, special education, HCI, or AI in education. I’m not looking simply for someone to evaluate a finished product. I’m interested in a collaborator who wants to investigate where that boundary should be and how we might design and study technology that helps struggling writers without replacing their thinking.
writing development, learning disabilities, AI in education, learning sciences, writing cognition, assistive technology, productive struggle, special education, language-based learning disabilities, human-AI interaction
