Enhancing Digital Reminder Systems for Dementia Care: Exploring Generative AI for Task Verification and Caregiver Support Through Co‐Design
Bibliographic record
Abstract
BACKGROUND: Caregivers of people living with dementia (PLwD) face significant stress, exacerbated by the challenges of verifying task completion despite the use of digital reminder systems. Generative AI, such as GPT, offers a potential solution by improving task verification and supporting caregiver decision-making. This feasibility study evaluated an AI-powered task verification system integrated with a digital reminder framework for PLwD. It explored (1) whether GPT can generate high-quality follow-up questions tailored to PLwD through few-shot prompting, (2) the accuracy of the system in flagging concerning responses, and (3) how to balance automation with caregiver control. METHOD: An anonymized dataset of 64 reminders was used to simulate interactions between caregivers, PLwD, and the AI system. GPT-generated follow-up questions were evaluated for quality with and without contextual information. A flagging mechanism classified responses as High, Medium, or Low concern, with critical categories including mealtime, personal safety, and daily hygiene. Simulated caregiver feedback was incorporated to refine question quality and system adaptability to keep AI-generated interactions relevant, clear, and minimally intrusive. Additionally, two EPLED (Engagement of People with Lived Experience of Dementia) members were involved in the design process, providing insights into the evaluation methods, few-shot examples, and usability of the system's responses. RESULT: Contextual information and caregiver feedback improved follow-up question clarity, specificity, and relevance, reducing ambiguity in PLwD responses. The flagging mechanism demonstrated high accuracy, particularly for safety-critical tasks such as medication and fall prevention reminders, while non-urgent tasks, such as leisure activities, presented greater subjectivity. Simulated caregiver feedback and EPLED members' input played a crucial role in adapting the system to individual needs and evaluating whether it genuinely reduces caregiver stress or simply redistributes their responsibilities. CONCLUSION: This study demonstrates the feasibility of integrating generative AI into dementia care. Context, caregiver input, and EPLED members' perspectives significantly improved task verification and decision support. By enhancing reminder effectiveness and refining caregiver alerts, AI-assisted verification has the potential to reduce caregiver stress and improve PLwD support. Future research should focus on real-world validation, user-centered customization, and scalability to optimize caregiver workload reduction and long-term adoption in home care settings.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.032 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".