The Elephant in the Easy Read Room: Can AI Really Do Easy Read?
Let’s talk about the big, awkward question: Can AI turn formal documents into Easy Read? The answer is tricky. It’s not a simple yes or no—it’s more like no, but also yes, sort of. AI isn’t about to take over the entire Easy Read process, but it does have a role to play, especially at the preliminary transcription stage. Let me explain.
Where AI Shines: The Preliminary Transcription Stage
Here’s the deal: when you’re creating Easy Read materials, the first thing you have to do is strip out all the jargon, complexity, and unnecessary fluff from the original document. This isn’t the final Easy Read version—it’s just a simplified, jargon-free starting point for the co-production process. Think of this Transcription Document as the scaffolding for the real work that comes next.
Traditionally, this job is done by a seasoned co-production workshop coordinator—someone who knows exactly what will work for a panel of Experts by Experience. They know how to “translate” complex text into something the panel can engage with, without overwhelming them or losing the core message.
But AI? AI is surprisingly good at this step. It can crank out a simplified, jargon-free version of a document in seconds, saving hours of time. And, let’s face it, there’s no point giving the original document to the panel anyway—it’ll be too dense and difficult for them to work with. A good transcription makes the whole workshop process smoother and more productive.
Where AI Falls Short: The Co-Production Process
Here’s where things get interesting. The transcription stage is just step one. The real magic happens when you bring in the panel of Experts by Experience to review, revise, and shape the text into something that truly works for Easy Read audiences. This step introduces a level of rigour that AI simply cannot replicate.
Why? Because AI doesn’t have a “lexical impediments model”—it doesn’t understand what an Easy Read target reader might struggle with. Sure, it can simplify text, but it doesn’t know what makes something truly accessible. It can’t pick up on nuances, like how certain words or phrases might still be confusing even if they’re technically simple. It doesn’t know what’s too abstract, what feels patronizing, or what’s missing altogether.
This is why the co-production workshop is irreplaceable. During the workshop, the coordinator leads the panel through the transcription, asking them to:
Read and Review: Does the text make sense to them?
Revise: What needs to change to make it clearer or more accurate?
Confirm Understanding: What did they understand from the text? If their understanding is wrong, the coordinator steps in to clarify the intended meaning and work through revisions.
This process isn’t just about simplifying language—it’s about making sure the information is actually usable. AI might produce something that looks like Easy Read, but without the insight of the panel, it’s never going to do the job properly.
AI and Easy Read Pictures: A Can of Even Wrigglier Worms
I’ll tackle this in more depth in my next blog post, but let me give you a sneak peek: AI has even bigger limitations when it comes to selecting or creating the pictures that accompany Easy Read materials. This step is just as crucial as the text, and getting it wrong can completely undermine the whole document.
The Bottom Line
AI is a tool—not a replacement for the co-production process. It can save time in the preliminary transcription stage, but it will never replace the insight, understanding, and rigour of a panel of Experts by Experience. The human element is what makes Easy Read truly accessible. Without it, you’re left with something that might look the part but doesn’t actually work for the people it’s meant to serve.
So, no, AI isn’t “good” at turning documents into Easy Read. But yes, it can be useful. It’s not the answer—it’s a stepping stone. And if we’re serious about accessibility, we need to use it wisely, without losing sight of what really matters: the people we’re creating these materials for.