AI That Moves With You: A Review of Interactive Technologies Powered by Large Foundation Models for Mobility Impairment
This is a strong synthesis paper that gives CHI a usable map of an emerging area: FM-enabled assistive interaction for mobility impairment. Its contribution is not a new system but a structured design-space view, corpus, and agenda grounded in a clearly delimited review of 26 papers and their recurring technical, evaluative, and ethical patterns.
Axes Lens
Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.
Contribution shape
- Knowledge form
- descriptive knowledge typical · 92/268
- Novelty type
- synthesis typical · 16/268
- Abstraction level
- field typical · 41/268
- Generalization target
- field argument typical · 55/268
- Validation mode
- survey synthesis typical · 10/268
Evidence profile
- Evidence strength
- strong typical · 158/268
- Claim alignment
- strong typical · 231/268
- Overclaim risk
- low typical · 53/268
Review Summary
This paper is most valuable as field-building infrastructure. Rather than claiming a breakthrough assistive system, it assembles and interprets a young, fragmented body of work on foundation-model-enabled interactive technologies for mobility impairment. The review makes a persuasive case that the important shift is from narrow, hand-crafted assistive pipelines toward systems that can combine language, vision, reasoning, and multimodal interaction in more adaptive ways. That framing matters for CHI because it connects rapid FM progress to concrete accessibility and assistive-technology questions instead of treating accessibility as a side application. The evidence base is appropriate for that kind of contribution. The authors report a five-database search with 6,249 screened records and a final corpus of 26 papers, then synthesize study design, evaluation approaches, FM techniques, integration patterns, interaction paradigms, and mobility contexts. The paper appears strongest when it stays descriptive and synthetic: it identifies where the literature is growing, what kinds of systems are being built, and what recurring tensions appear around safety, privacy, personalization, trust, and ecological validity. It also usefully broadens mobility impairment beyond a narrow clinical category by including mobility-relevant sensory and neurological contexts. The main caution is that the review inherits the weaknesses of the underlying literature. The paper itself notes small participant groups, short-term lab studies, proxy tasks, limited participant diversity, and evaluation subjectivity or bias in some reviewed systems. So the agenda it proposes should be read as a disciplined synthesis of an emerging area, not as settled evidence that FM-based assistive systems are already robust in real-world use. Even with that limitation, the paper is a meaningful CHI contribution because it gives researchers a clearer vocabulary, scope, and set of priorities for future work.
What Changed
Canon before
Traditional interactive assistive technologies for mobility impairments rely on narrow, hand-crafted pipelines or rigid input-output modalities, limiting flexibility, adaptability, and user autonomy. Existing AI and accessibility research mostly address individual disability groups or narrow technical tasks without a unified design space for large foundation models (FMs) in interactive mobility asistive systems.
Departure from common sense
The paper argues that assistive technologies for mobility impairments need not remain narrow, hand-crafted, or modality-rigid. Instead, foundation models can generalize across tasks and support more fluid multimodal interaction, shifting the field toward adaptive systems that better match varied user needs and contexts.
Actual novelty
The paper's main contribution is a scoping review and synthesis of 26 full papers on FM-enabled interactive technologies for people with mobility impairments, paired with a conceptualization of FM-enabled interactions as a design space, a tabulated corpus with reproducible codebook, and a forward-looking research agenda.
Evidence
The paper supports its claims through a scoping review spanning five databases and 6,249 screened records, resulting in a final corpus of 26 papers. The evidence is appropriate for a synthesis contribution: the authors summarize study designs, evaluation approaches, FM techniques, integration patterns, interaction paradigms, and impairment contexts, then derive cross-cutting technical, ethical, and methodological observations. The strongest support is for descriptive mapping and agenda-setting rather than causal claims.
“ future research topics. We contribute: (i) a conceptualization of FM-enabled interactions for mobility impairment functioning as a design space; (ii) a tabulated corpus with a reproducible codebook; and (iii) a forward agenda to guide and inspire the design of future mobility-assistance interactive systems within human-computer interaction (HCI) and CHI community”
actual novelty · Abstract · confidence 0.98
“– have sought to bridge physical limitations by creating alternative pathways for communication and action [104, 120]. While these systems have provided crucial support, they have traditionally relied on narrow, hand-crafted pipelines or rigid input–output modalities, limiting flexibility, adaptability, and user autonomy [92]”
departure from common sense · 1 Introduction · confidence 0.96
“ Evaluations such as Patrika, Talk2Care, WorldScribe, and GenAssist typically involved small participant groups and short-term, lab-based testing, which restricted the generalizability of findings”
limitation · E Study Limitations · confidence 0.98
“ Searching five databases, we screened 6,249 records and included 26 full papers. We first summerize descriptive results including study design and evaluation approaches of the reviewed studies”
validation scope · Abstract · confidence 0.98
Limits
Method limits
The review is bounded by its corpus and coding choices. The paper synthesizes 26 papers from 2022 to 2025 and notes that reviewed studies often relied on small participant groups, short-term lab testing, proxy tasks, and subjective or potentially biased evaluation procedures, which constrains generalizability and ecological validity of the synthesized evidence.
Deployment limits
As a review paper, it does not deploy a new system itself. Its practical implications depend on the maturity of the underlying literature, which the paper notes is dominated by short-term and lab-based evaluations, with only a few longer in-the-wild efforts.
Boundary conditions
The claims apply to FM-enabled interactive technologies for mobility impairment contexts as defined functionally by the authors, including motor impairments and mobility-relevant sensory or neurological contexts. The contribution is primarily a synthesis of an emerging literature rather than a definitive account of all assistive AI systems or all disability domains.
Position in field
This paper helps consolidate an emerging HCI area at the intersection of foundation models, accessibility, and assistive technology. Its value lies less in introducing a single new artifact than in organizing scattered prototypes, clarifying recurring design and evaluation patterns, and framing future work around robustness, personalization, safety, and equity.