Progress Over Perfection: Reframing How the Deaf Community Engages with AI

ASL Translation

NZSL Translation

AI is everywhere these days - answering questions, translating languages, even helping people communicate across barriers. For many hearing users, AI tools have been around for a while now. Systems like Siri, Alexa, and ChatGPT are already part of daily life. They’re understood as “works in progress” sometimes robotic, sometimes wrong, but always improving. 

For the Deaf community, though, the relationship with AI often looks different. When it comes to digital signing or language accessibility, expectations are higher and the technology is still evolving, still learning.

There’s a strong desire for tools to get it right, to sign fluently, accurately, and respectfully. That expectation comes from a good place: deep care for language integrity and Deaf identity.

But it can also make it harder to accept emerging versions of AI technology, even when those tools are designed with us and for us. (us being the Deaf community) 

Why Mistakes Matter 

It can feel uncomfortable to see AI sign incorrectly, use the wrong facial expression, choose a literal sign instead of a descriptive one, or fumble the rhythm of a signed sentence. For a community that has fought for centuries to have sign language recognised and respected, “almost right” can feel like another form of misrepresentation. 

But AI learns the same way people do, through trial, feedback, and time. Every imperfection is an opportunity for improvement. That’s how hearing users experience voice-based AI: they laugh when Siri misunderstands a command or when Alexa mishears a name, because they know it’s a work in progress, not a finished product. They understand that progress comes through use, not avoidance. 

When it comes to accessible AI, the Deaf community can adopt that same mindset in our own engagement. Instead of seeing mistakes as failure and shutting innovation down - we can recognise those flaws as part of the process: evidence that the system is learning, and that our feedback truly matters. 

The Power of Feedback 

Every time a Deaf user provides feedback, whether that’s pointing out a mistranslated sign, suggesting smoother transitions, or clarifying a facial cue that information helps AI evolve. Behind every update is a community voice saying, “Here’s how to make this better.” 

This kind of collaboration is what makes accessible technology genuinely inclusive. It’s not about waiting for perfect AI signing before we engage. It’s about co-creating a tool that becomes more authentic over time, through safe and thoughtful use.

Progress happens fastest when we, as Deaf people, participate not when we stand back and wait for perfection to happen around us. 

Patience as a Form of Empowerment 

It’s tempting to wish for instant results, a signing avatar that looks and feels entirely natural from day one. But technology doesn’t grow that way. The path from prototype to polished is paved with trial, testing, and trust. Accepting imperfection isn’t lowering standards; it’s recognising that our active involvement in the process is what raises the bar. 

When we approach accessible AI with patience, we give it room to develop into something that truly serves us, led by us. We allow developers to learn, adjust, and improve, with Deaf voices guiding every step. 

A Shared Journey 

Accessible AI isn’t just about technology. It’s about relationships - between developers, linguists, interpreters, and the Deaf community itself. Each group brings unique knowledge, and progress depends on listening to one another. 

When we view AI as a living, learning system rather than a finished product, we create space for shared ownership. The goal isn’t “flawless signing” it’s continuous growth, rooted in community collaboration and cultural respect. 

Mistakes, in this context, aren’t setbacks. They’re signs of movement - proof that AI is learning to reflect real human language: nuanced, expressive, and evolving. 

At Kara, we see this learning process as a collaboration - between technology and community, where feedback drives growth and accessibility evolves together. 

Looking Ahead 

The future of accessible AI will be shaped by how we engage with it now. If we expect perfection, we risk slowing progress. If we embrace the process, mistakes and all, we open the door to technology that reflects who we are and how we communicate. 

Just as Siri and ChatGPT have grown through millions of hearing interactions, AI signing tools will improve through Deaf-led feedback. Every correction, every suggestion, every moment of patience is a step toward better representation. 

Progress in accessibility doesn’t come from flawless technology. It comes from community trust, feedback, and a shared belief that “good enough” today can lead to extraordinary tomorrow. 

"If AI grows through its flaws, how might you grow with it?” - Grace Covey