Access-Stabilising Support

A design framing, introduced by Bhuiyan et al. (2026), that positions AI in Deaf education not as an autonomous translator or replacement instructor but as a mediated tool whose role is to preserve visual access, reinforce teacher-validated signs, and sustain comprehension across classroom disruptions. Access-stabilising support operates under teacher authority and community governance rather than model-centric autonomy, with features such as slow and expressive signing, adjustable pacing and replay, regionally validated vocabulary, offline-first resilience, graceful degradation under intermittent connectivity, and private clarification pathways that protect emotional safety. The framing shifts design priorities from translation accuracy on benchmark corpora to continuity maintenance under structural instability.