Artificial intelligence (AI) is rapidly transforming health care, including dental practice, through innovations like automated caries detection and predictive treatment planning. ^, However, for the estimated 61 million adults with disabilities in the United States, these advances often remain inaccessible due to disparities in oral health and care. Traditional AI development has focused on efficiency and scalability, often overlooking the complex social, economic, and structural determinants that shape health outcomes. As a result, AI tools risk perpetuating existing health disparities when designed primarily for mainstream populations.

This article proposes prosocial AI—a framework that envisions AI not merely as a tool for optimization but to actively promote equity, inclusion, and human dignity. Evidence shows that individuals with intellectual, physical, sensory, and psychiatric disabilities experience significantly higher rates of periodontal disease, tooth loss, and untreated decay, stemming from systemic exclusion, inaccessible facilities, insufficient provider training, and discriminatory attitudes. ^,,

These disparities have far-reaching impacts, compromising nutrition, social participation, employment, and overall quality of life. Caregivers shoulder much of this burden, often without sufficient support or training. While traditional approaches focus on expanding access to existing services, they often fail to address deeper structural barriers. Prosocial AI prioritizes whose needs are central and whose well-being is promoted, valuing the expertise of people with disabilities and their caregivers in navigating complex health systems.

Potential applications include diagnostic tools adapted to diverse communication styles, personalized educational platforms, teledentistry to reduce travel barriers, and AI-driven caregiver support. This article explores these possibilities, analyzing disparities, evaluating current AI through an equity lens, proposing a prosocial AI development framework, and showcasing the Compassionate Caregiver Companion Coach (C4).

Oral health disparities among people with disabilities remain one of modern health care’s most persistent inequities, despite decades of advocacy and policy efforts. Individuals with disabilities experience worse oral health outcomes across the lifespan, with disparities beginning in childhood and compounding over time.

Survey data highlight the stark reality: adults with disabilities are twice as likely to have untreated tooth decay and thrice as likely to be edentulous compared to adults without disabilities. Among adults aged 18 to 64 years, severe periodontal disease affects 27.2% of those with disabilities versus only 9.8% of those without. Individuals with intellectual disabilities face particularly severe challenges, experiencing high rates of undetected periodontal disease, tooth loss, and untreated caries. Physical disabilities introduce barriers related to mobility, positioning, and dexterity. Sensory impairments can complicate communication and health literacy, further hindering oral health maintenance. People with multiple disabilities or complex medical conditions often experience even greater challenges, as increasing disability severity correlates with worsening oral health.

Disparities are fueled by interlocking barriers at individual, provider, and system levels.

• •

  Physical and architectural barriers: Many dental offices remain physically inaccessible. Only 37% provide accessible parking, and fewer than 20% offer adjustable chairs or lifts. Standard dental equipment often fails to accommodate diverse body types or positioning needs, leading to suboptimal care.

• •

  Provider training and comfort: Many dental providers feel unprepared to care for patients with disabilities, particularly individuals with intellectual disabilities or autism. Inadequate training results in shorter, crisis-driven visits, less thorough examinations, and reduced patient education.

• •

  Communication and behavioral challenges: Effective dental care requires nuanced communication that traditional methods often lack for individuals with intellectual disabilities, autism, or communication disorders. Behavioral challenges, such as anxiety or sensory overload, can further complicate treatment without appropriate support strategies.

• •

  Systemic and economic barriers: Financial obstacles persist, including limited dental insurance, minimal Medicaid dental benefits for adults, and low reimbursement rates for treating patients with disabilities. Additionally, few providers accept Medicaid, severely restricting options for patients who rely on it.

Family caregivers and direct support professionals play an essential role in maintaining oral health for people with disabilities. Their knowledge, skills, and attitudes influence outcomes. However, caregivers often lack adequate training, leading to knowledge gaps and inconsistent oral hygiene practices. High caregiver stress and burnout can further compromise daily routines, contributing to the cycle of poor oral health.

Although AI innovations in dentistry have advanced rapidly, existing tools often fail to address the specific needs of people with disabilities. Most AI applications focus on diagnostic imaging, treatment planning, and practice management that benefit dental professionals rather than patients with complex needs. Diagnostic AI systems typically depend on high-quality images and standardized protocols, which can be difficult to achieve for patients with movement disorders or behavioral challenges. These systems often lack contextual understanding of disability-related factors that impact care. Treatment-planning AI rarely accommodates adaptations necessary for patients with disabilities. Critically, disability communities remain largely absent from the development and evaluation of dental AI, contributing to the limited relevance and accessibility of these tools. This underscores the urgent need for AI development approaches that prioritize equity, inclusion, and community engagement from the outset.

Creating AI systems that effectively serve marginalized communities demands a fundamental shift in how health technologies are designed and evaluated. A human-centered approach emphasizes the unique needs and strengths of individuals—both patients and caregivers—and prioritizes social impact, equity, and human flourishing over narrow technical performance metrics.

Prosocial AI is developed with deliberate intent to bring out the best in people and society. Key principles include

• •

  Accessibility-first design: Technology should consider diverse abilities, communication styles, and user preferences from the outset. Unlike designing for the “average,” prosocial AI prioritizes users with the most complex needs, which often leads to solutions benefiting everyone. ^,

• •

  Community participation and codesign: Rooted in the principle “Nothing about us without us,” meaningful involvement of people with disabilities and their caregivers ensures that AI solutions address real needs and respect community strengths. Such participation requires sustained engagement, fair compensation for community expertise, and shared power in decision-making.

• •

  Equity over efficiency: Traditional AI development emphasizes scalability and efficiency, which can overlook the need for personalized solutions. Prosocial AI explicitly prioritizes equity, recognizing that fair outcomes might demand customized approaches (eg, diagnostic tools that gather data from caregivers when patients cannot communicate directly).

• •

  Transparency and explainability: Users must understand how AI systems reach decisions affecting their lives. Prosocial AI should clearly communicate its reasoning, limitations, and uncertainties in accessible formats.

While AI holds tremendous promise, its benefits will remain limited without intentional design for inclusion. Key applications for prosocial AI in oral health care include

• •

  Diagnostic support systems: Traditional AI diagnostics rely on standardized imaging protocols, which can be problematic for patients with movement disorders or behavioral challenges. Prosocial diagnostic AI would accept diverse data inputs—such as caregiver photographs or verbal descriptions—and train algorithms to interpret varied information, including medication side effects or the use of adaptive equipment.

• •

  Teledentistry and remote monitoring: Although teledentistry can reduce barriers, current platforms often assume standard digital access and literacy. Prosocial teledentistry AI would support alternative communication methods, simplified interfaces, and the ability to manage irregular data while focusing on trends and patient support.

• •

  Caregiver support and education platforms: AI-driven platforms can offer caregivers tailored guidance that adapts to specific patient needs, caregiver skills, and environmental conditions. Such systems could combine education with practical problem-solving while addressing caregiver stress and well-being.

• •

  Assistive technologies for direct patient support: AI could help individuals maintain oral health independence. Potential solutions include voice-activated brushing instructions, smart toothbrushes that adjust for motor impairments, communication aids for dental visits, and visual assistance tools for individuals with visual impairments.

Equally important as technological capabilities are the design processes behind AI systems, which must emphasize inclusivity and personalization.

• •

  Flexibility and customization: AI systems should be adaptable to the vast diversity within disability communities, offering user-controlled settings and intelligent defaults for personalized interaction.

• •

  Progressive disclosure and scaffolding: Information and features should be presented progressively, allowing users to access more complexity as needed and avoiding cognitive overload. ^,

• •

  Error prevention and recovery: Systems should anticipate user errors, offer gentle corrections instead of rigid error messages, and provide multiple pathways for completing tasks—an essential consideration for users with cognitive disabilities.

• •

  Cultural responsiveness: AI must respect cultural differences in health beliefs and practices. This requires diverse training data and adaptable algorithms to ensure solutions are relevant across communities.

Translating these principles into practice demands careful implementation and sustained community engagement. The next section illustrates how these concepts come together in the case of the C4.

The C4 was developed as a tool that serves as a “coach in the pocket” to guide, support, and uplift caregivers, while connecting them with their peers, local service providers, and quality information. The C4 is anchored in the multidimensional (Perspective; Optimization; Zenith; Exposure [POZE]) paradigm, which looks at human existence as a multidimensional composition, with aspirations, emotions, thoughts and sensations, and individuals as part of a social net. This multilevel dynamic is the cause and consequence of values that characterize the needs and preferences of both the caregiver and the patient. (Developed through intensive community engagement and iterative design, the tool focuses on empowering caregivers while maintaining patient autonomy and dignity. It exemplifies how prosocial AI principles could be applied in oral health care for individuals with disabilities.

C4’s development was rooted in thorough listening sessions with caregivers and self-advocates. Caregivers consistently expressed the need for practical guidance, emotional support, and solutions for daily challenges, particularly in managing oral care for individuals with autism or those living in group home settings. Simultaneously, self-advocates and parents emphasized the importance of preserving autonomy and dignity in personal care decisions. These insights shaped C4’s design to balance robust caregiver assistance with respect for individual independence.

C4 is a conversational AI accessible via various digital platforms, integrating multiple prosocial AI principles.

• •

  Empathetic conversational AI: C4 uses natural language processing refined through real-life communication patterns from experienced support workers. This training allows the AI to recognize caregiver emotions such as frustration or anxiety and respond with empathy. Its interface adapts dynamically to different literacy levels and communication styles, embodying human-centered design.

• •

  Personalized care planning: C4 creates individualized oral health plans based on a holistic assessment of patient-specific factors (including disability type, communication abilities, sensory sensitivities, and medication side effect) and caregiver characteristics such as experience level, stress indicators, and cultural context. This customization reflects the principle of flexibility and adaptation.

• •

  Real-time adaptive guidance: The AI offers immediate support during oral care routines, learning from ongoing interactions to refine recommendations. For instance, it might suggest specific strategies for an individual with autism based on recorded preferences, demonstrating how equity can be prioritized over efficiency through personalized support.

• •

  Integration with health care providers: C4 facilitates communication between caregivers and dental professionals by securely logging observations and assisting in symptom documentation. This transparency allows dental teams to better understand patient contexts while respecting privacy and caregiver control over information sharing.

• •

  Community curation and self-care: Recognizing the isolating nature of caregiving, C4 fosters connections among caregivers facing similar challenges. It also includes tools for caregivers to monitor and manage their own mental and physical health, promoting well-being alongside caregiving duties.

C4’s design aims to produce several outcomes.

• •

  Increased caregiver confidence and competence, with users reporting feeling more informed and less anxious.

• •

  Improved daily oral hygiene routines for individuals with disabilities, facilitated by tailored strategies and reduced care resistance.

• •

  Earlier identification of oral health issues, especially critical for nonverbal individuals.

• •

  Enhanced caregiver well-being, with reduced stress and a sense of support from both AI assistance and community connections.

• C4’s development highlighted key challenges relevant to broader prosocial AI implementation:

• •

  Balancing automation and human connection: Ensuring AI complements rather than replaces human judgment and relationships remains a delicate balance.

• •

  Privacy and data security: Users voiced concerns about potential misuse of sensitive data, necessitating robust transparency and privacy safeguards.

• •

  The digital divide: Differences in technology access and digital literacy pose significant barriers to equitable deployment.

• •

  Cultural and linguistic diversity: Broader engagement with diverse communities is essential to ensure C4’s cultural relevance and accessibility.

• •

  Health care system integration: Integrating new AI tools into existing health care infrastructures requires navigating technical complexities and organizational workflows.

Despite these challenges, C4 illustrates how prosocial AI can deliver personalized, meaningful support for both individuals with disabilities and their caregivers, providing an example for future innovations.

Developing prosocial AI in health care demands ethical considerations that go beyond traditional biomedical ethics, encompassing justice, participation, and social responsibility. For AI tools serving people with disabilities, these ethical questions are especially critical, given the historical exclusion and exploitation of these communities.

The disability rights movement’s principle “Nothing about us without us” remains fundamental. People with disabilities must be integral partners at every stage of AI development. Meaningful engagement involves shared decision-making, compensation for community expertise, and sustained collaboration. Practical strategies include forming community advisory boards, hiring people with disabilities as coresearchers, and ensuring developers remain responsive to community feedback. The C4 case study underscores the value of continuous engagement and flexible design to reflect evolving community needs.

Traditional research and development processes often reinforce power imbalances, with people with disabilities sometimes feeling pressured to express gratitude instead of voicing critiques. Ethical AI development requires actively dismantling these dynamics, fostering equitable participation, and ensuring community members genuinely influence decisions. Financial compensation is both an ethical obligation and a practical necessity for meaningful engagement.

Disability intersects with multiple social identities, including race, gender, class, and cultural background, creating layered experiences of discrimination. AI systems that ignore this intersectionality risk perpetuating multiple forms of bias. Cultural humility—an ongoing process of self-reflection and openness to learning—is essential for developers. This mindset influences AI development through diverse training data, cross-cultural testing, and adaptable interfaces, ensuring technology reflects the realities of varied communities.