Researchers have developed AI-Care, a conversational AI system designed to help individuals with Alzheimer's disease and related dementia manage everyday tasks through natural voice interaction.

The system addresses a critical gap in digital accessibility for people with cognitive impairments. Traditional digital tools like calendar apps require multiple steps that can overwhelm Alzheimer's patients, creating barriers to independent living.

AI-Care operates as a voice-first chatbot built on top of existing caregiving platforms. Users can set calendar reminders, organize to-do lists, and coordinate care tasks through simple spoken requests rather than navigating complex interfaces.

Technical Architecture and Safety Controls

The system uses LangGraph-based stateful orchestration to process requests through multiple stages: sanitization, intent classification, context loading, safety checks, slot collection, tool execution, and response composition.

Safety-critical responses around medications and allergies are grounded in caregiver-verified records rather than AI-generated content. The system explicitly avoids making autonomous medical or treatment decisions.

Incomplete requests trigger controlled multi-turn clarification rather than guessing or silent failure. Voice output uses ElevenLabs text-to-speech technology, with longer responses chunked to prevent rushed playback.

A preliminary pilot with four individuals with mild-to-moderate Alzheimer's showed positive results. Participants found the system trustworthy, competent, and likable while successfully completing coordination tasks through conversation.

The research team includes eight authors from multiple institutions, led by Preyash Yadav and Michelle Cohn. Their work was published on arXiv in May 2026.

The system represents a growing trend toward accessible AI interfaces for vulnerable populations, focusing on reducing cognitive load rather than adding features.