When I first started coaching my dyslexic son I was excited to introduce him to various speech-to-text technologies. I believed these tools would be a complete solution to his struggles with writing. I thought he would quickly embrace this technology, but to my surprise, he resisted. Despite investing in software that could transcribe his spoken words, Judah was not interested.
Initially, I assumed he wanted to see his thoughts concretely on paper. So, we focused on making notes and studying through handwriting, even if it was untidy and misspelled. However, his resistance persisted over the months.
I discussed this with Judah and his cousin Jayce, my two favourite dyslexic teens and we even talked about it in a clip you can view here.
In my interview with Nick Koshnick of vnote.ai for the Wonderfully Wired podcast this month, we delved into the limitations of speech-to-text technology. Nick explained that while speech-to-text was a necessary step towards developing large language models and AI, it actually set us back in some ways. Instead of simplifying the writing process, it made it more complicated. Dyslexic individuals now had to sift through their rambling thoughts on paper, making it even harder to find clarity.
What AI can do for a dyslexic thinker is more than just transcribing speech. It can use large language models and question-based approaches to guide the writer towards clarity. This process, which Carolyn Maguire calls "holding up the metacognition mirror," is what a good teacher and active listener knows to provide someone learning to write well. It helps the thinker work through their rambling thoughts to achieve clarity.
Nick emphasized that voice is greater than speech, and writing is more than just text. We need a model that doesn't simply convert speech to text but allows a voice to explore interconnected reasoning and achieve clear and concise communication: writing.
In summary, while speech-to-text technology was a necessary step towards AI-powered tools, it is not the solution we assumed it would be for dyslexic students. Instead of eliminating the need to write or type, it adds another layer of complexity. Dyslexic thinkers need tools that help them funnel their interconnected ideas into clear and simple communication, not just transcribe their rambling thoughts.
Core Message:Â Speech-to-text technology, while a necessary step towards AI-powered tools, is actually a barrier for dyslexic students rather than the accommodation we assumed it would be.
Impact on Dyslexic Students:Â Dyslexic thinkers struggle to turn their many interconnected ideas into clear and simple communication. Speech-to-text technology adds another layer of complexity, making it harder for them to achieve clarity.
Comentários