The Transformative Power of AI: Giving Deaf and Hard-of-Hearing Americans New Ways to Connect

This week is International Week of the Deaf. According to the World Health Organization, roughly 466 million people worldwide live with disabling hearing loss, and for many of them, something as ordinary as a doctor’s appointment, a work meeting, or a phone call with a loved one has required an interpreter, a favor from a hearing friend, or simply going without. With the growing investment in artificial intelligence (AI), that gap is closing faster than most people realize. For decades, the tools available to deaf and hard-of-hearing people were valuable but scarce: certified sign language interpreters, captioners, and specialized hearing aids that are expensive or simply unavailable. A rural patient couldn’t always get a live interpreter for a same-day medical visit. A small business couldn’t always afford real-time captioning for a staff meeting. The technology existed to help, but the access to it could not scale with the need. AI is helping to expand access by putting real-time communication support in everyone’s pocket, on their face, and in their ears — on demand, at a fraction of the historical cost.  Modern hearing aids no longer just amplify sound, they can adapt to it. Starkey’s Genesis AI platform uses onboard machine learning to continuously analyze a wearer’s acoustic environment and adjust in real time, making as many as 80 million processing adjustments per hour to isolate speech from background noise. In published research, the company found the technology improved speech understanding in noisy environments by 30% compared to going unaided. Starkey’s Omega AI models its noise processing on how different regions of the human auditory cortex work together, and testing found that compared to previous models, it delivered a further 28% gain in speech intelligibility.  That single statistic represents something profound: a noisy restaurant, a crowded classroom, or a family holiday gathering, once the hardest environments for someone with hearing loss to participate in, has become far more navigable. Automatic speech recognition has also made real-time captioning genuinely real-time. AI-powered captioning glasses now capture a speaker’s voice through built-in microphones, process it through cloud- or on-device speech-to-text engines, and project transcribed text directly into the wearer’s field of vision within moments, letting a deaf or hard-of-hearing person follow a conversation without delay. The same underlying technology now powers live captions in video meetings, at conferences, and in classrooms. Zoom, for instance, has rolled out AI-driven speech modeling, developed in part with researchers at the University of Washington, which has made real-time captions more accurate for deaf and hard-of-hearing participants, supporting over 36 languages. AI is also working in the other direction — turning spoken and written language into sign language, and sign language into text. Signapse.ai generates photo-realistic AI sign language video from text and is already deployed to translate time-critical travel announcements in UK rail stations and airports, helping deaf sign-language users who are otherwise excluded from spoken-word announcements. Meanwhile, SignAll is the most advanced automated ASL-to-English camera-based translation system, aimed at the reverse challenge: helping hearing people understand signers in real time. Innovation-killing red tape aimed at broad AI systems, however well-intentioned, risks slowing exactly the kind of rapid, competitive product development that has driven these gains for the deaf and hard-of-hearing community. Policymakers should take note of what a free-to-innovate marketplace has already delivered, and make sure the next generation of AI-powered accessibility tools has the room to grow and help even more Americans.