Why might many patients still prefer to see a human doctor, even as AI systems become better at answering certain medical information questions? The reason goes beyond medical knowledge or correct answers. Healthcare support involves communication that resolves many forms of uncertainty: people’s needs, routines, abilities, and environments vary widely, and interactions must adapt to these contexts to provide appropriate support. Today’s context-aware AI systems still struggle with such uncertainty. For example, they may misinterpret user behavior and confidently suggest irrelevant actions, leaving users to correct the system while unsure whether it will adapt in future interactions.
This autoethnographic study by a Deaf researcher with a multicultural background examined the use of automatic speech recognition and text-to-speech (TTS) on mobile/computerdevices and augmented reality (AR) smart glasses to support communication for Deaf and Hard of Hearing individuals in educational and everyday contexts. Mobile and computer-based tools enable active participation through larger displays but require sustained screen attention, reducing access to facial cues. AR smart glasses support hands-free transcription and visual attention but lack TTS and bidirectional interaction. The findings suggest that improving comfort, reliability, and adaptive input methods is essential, and that combining AR smart glasses with mobile and computer tools offers complementary support for accessibility and social inclusion rather than replacing human communication.
