When Assistive Technologies become Provocations: Unpacking Access in HCI practices using Crip Technoscience, Mouth Interfaces, and XR
Puneet Jain, Ayush Sharma, Sidharth Chaudhary, Vivek Rawat, Akhilesh Kumar Bhagat, Kratika Jain, Christian Bayerlein, Christopher Lloyd Salter, Gowdham Prabhakar · 2026 · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI '26) · doi:10.1145/3772318.3791535
Summary
This CHI 2026 paper reframes assistive technology as political provocation rather than technical fix. Developed in long-term collaboration with two disabled artists - Eric Desrosiers (muscular dystrophy) and Christian Bayerlein (spinal muscular atrophy), both mouth-operated joystick users - the authors built a camera-based mouth interface for Meta Quest XR headsets. The device classifies tongue, cheek, jaw, and lip gestures (42 detected via a Project Babble-derived ML pipeline) to drive navigation in VR, directly challenging the hand-tracking defaults Meta positions as universal. Rather than evaluating the interface as an accessibility add-on, the team exhibited it as XR art-activism in public galleries and designed a study using what they call the narrative-flip method: 14 early-stage HCI master's students (ages 21-26) performed a 5-minute Unity cube-collection task using four mouth gestures, gave a pre-reveal semi-structured interview, then watched a disclosure presentation revealing the interface's disability-led, activist origins, and finally gave a post-reveal interview. The paper is explicitly grounded in crip technoscience (Hamraie and Fritsch), critical disability studies, and the disability justice tradition. Drawing on inductive reflexive thematic analysis of roughly nine hours of transcripts - re-interpreted through a debrief with co-author Bayerlein - the authors organise findings into four clusters: Bodies/Norms/Un(Comfort), Solutionist Drift, Empathy and its Fragility, and Access as More than a Solution. Methodological limitations (demand characteristics, transferability, risk of disability-simulation misreading) are discussed candidly.
Key findings
Pre-reveal, participants consistently framed the mouth interface through a solutionist lens - imagining it for 'emergency scenarios,' amputees, or drivers 'who cannot use controllers' - and described the experience as awkward, embarrassing, or physically straining after five minutes. Body awareness shifted visibly: participants reported 'forgetting their hands' and becoming hyper-aware of mouth pressure, saliva, and social judgement around sticking out the tongue. Post-reveal, participants' framings shifted along three axes the authors label as access-as-fix, access-as-political, and access-as-agency. Empathy rose sharply, but the team documents an empathy fragility: participants slid back into product-thinking, proposing commercialisation ('launch it as a full-fledged product'), BCI integration, or video-game repurposing - re-absorbing the provocation into market logics. Co-author Bayerlein's debrief reframed 'awkwardness' as a cultural verdict against non-normative bodies rather than an interface flaw, and critiqued empathy when it operates as pity rather than co-creation. Only a handful of participants referenced the disabled collaborators' actual choices when proposing 'improvements,' illustrating what Shinohara and Wobbrock call the shadow of misperception and echoing Jackson's disability-dongle critique.
Relevance
For accessibility practitioners, the paper is a sharp corrective to checklist and 'fix' mentalities: it shows that even HCI students exposed to an explicit activist framing revert to solutionist defaults within minutes, which has direct implications for inclusive-design curricula, procurement conversations, and the way teams treat disabled collaborators (as co-designers versus 'users' or 'testers'). The narrative-flip method itself is transferable - teams could use delayed-disclosure exercises in workshops to surface ableist assumptions before product decisions harden. Limitations are real: a small sample of design-school students from a single Indian institution limits transferability, and the study deliberately avoids usability metrics, so practitioners looking for XR input benchmarks will not find them here. Still, the work is an important signal that XR's hand-centric defaults are a live accessibility issue and that disabled artists' everyday practices are a generative - not remedial - source for interaction design.
Tags: crip technoscience · extended reality · XR accessibility · mouth interface · provocation · assistive technology · disability-led design · ableism · HCI research methods · activist design · participatory design · disability justice · critical disability studies