When we think about communication, it’s not just about what we say, it’s how we say it. Our tone, humor, rhythm, and even choice of words shape how others perceive us. But what happens when someone uses an AAC (Augmentative and Alternative Communication) system to speak? And now that AI is stepping in to help generate full messages, will we all start sounding the same?
Let’s talk about voice identity, and why it’s becoming one of the most important (and overlooked) frontiers in AAC.
The Problem with Generic
Many AAC tools focus on being efficient and grammatically correct – which is great. But here’s the thing: most people don’t speak like a textbook.
We joke. We’re sarcastic. We ramble. We repeat ourselves. We soften or sharpen our tone based on who we’re talking to.
If an AI-powered AAC system expands a thought into a polished, formal sentence every time, something gets lost: authenticity
The Gap Between Saying Something and Sounding Like Yourself
Traditional AAC systems focus on functional communication: requests, needs, basic exchanges. That’s essential. But users have always wanted more. They want to sound like themselves, not like a default robot voice or a textbook sentence generator.
Now, with the rise of AI-based sentence expansion tools, there’s a new opportunity: helping AAC users move beyond generic phrasing to something that sounds a bit more… them.
But here’s the catch:
If everyone uses the same AI, will we all start to sound alike?
Research already shows this concern is real. In studies where AAC users were offered AI-generated sentence suggestions, users loved the convenience, but only when the output reflected their personal style. If it didn’t feel authentic, many preferred to retype or reject the suggestion altogether.
What Voice Identity Really Looks Like
Because for AAC users, like everyone else, communication isn’t just about what you say – it’s about how you say it.
That might mean:
- Saying “Nope” instead of “I would prefer not to.”
- Sounding cheerful in one conversation and short and direct in another.
- Making jokes. Teasing. Being a little sarcastic. Or just being casual.
What users say is only half the story; how they say it is the rest.
Introducing the Idea of Personas
Some emerging tools are now exploring ways to let AAC users (or their teams) set a communication persona. This simple shift can make sentence suggestions feel more natural, more expressive, and more like the user.
Let’s take a simple message:
“I don’t want to go.”
With voice personalization, that could become:
- Formal: “I would prefer not to attend that event today.”
- Playful: “Nope, not in the mood for a crowd today!”
- Direct: “I’m not going.”
- Empathetic: “I’m feeling off today, I think I’ll skip it.”
These shifts in tone aren’t just cosmetic. They reflect who the communicator is. They support identity, intent, and connection—things that matter just as much as vocabulary or grammar.
Looking Ahead
As AI becomes more embedded in AAC, we need to prioritize voice identity, not just communication access. It’s not enough to help someone say something. We need to help them say it their way.
Because in the end, a voice that doesn’t feel like yours… doesn’t feel like communication.
Want to see it in action?
Check out our demo video to see how sentence expansion works in SwiftSpeak.
Get in touch if you’d like to explore how you or someone you support would like to try the beta version and help shape what comes next.
