AI Augments Front Line Care: Paperwork Reduced by 30%, Nighttime Falls Decline
High turnover among social workers—whose tenure can range from just three months to over ten years—means many are thrust into demanding care settings before fully acclimating to their roles. In 2024, Taiwan’s long-term care sector faced a shortfall of more than 4,000 workers. Under tight schedules, social workers conducting home visits must rapidly assess clients’ needs across more than 140 long-term care service items, while completing extensive paper-based records and care recommendations. The administrative process alone can take two to three hours, creating significant pressure for new entrants driven by strong commitment but limited experience.
To address the problem of frontline overload and the difficulty of transferring institutional knowledge, III-STI developed the Long-Term Care AI Agent. During home visits, the system captures natural speech, automatically transcribes and interprets conversations, generates care summaries, and provides real-time service recommendations such as meal delivery, material assistance, or assistive device referrals. Acting as a virtual senior care consultant, the AI Agent reduces documentation time by over 30 percent, enabling new caregivers to quickly grasp case details and improve decision-making efficiency. Pilot deployments with local health bureaus and private care providers have yielded positive feedback on efficiency and usability, demonstrating strong potential to relieve frontline pressures.
For care facilities where nighttime staffing is particularly constrained, III-STI partnered with Compal Electronics and Humetrics to integrate AI with 5G communications and smart mattresses, developing the 5G- and AI-enabled Smart Mattress Care System. The system provides real-time monitoring of older adults’ sleep and activity, predicts bed-exit risks up to 90 seconds in advance, and issues automated alerts to support timely intervention by caregivers. The solution not only helps prevent falls but also significantly reduces reliance on labor-intensive, high-frequency night rounds. Already deployed in multiple care facilities, the system has served more than 6,000 person-times, markedly enhancing care safety and improving working conditions. In recognition of its innovation, the system received a 2024 R&D 100 Award, showcasing Taiwan's technological soft power in the fields of smart care and AI-driven communications integration and application.
AI Expands the Boundaries of Care, Enhancing Rehabilitation Diagnostics and Safeguarding Age-Friendly Mobility
Beyond frontline care settings, the application of AI in rehabilitation medicine and public safety has become increasingly urgent. In the rehabilitation field, traditional gait and posture assessments typically rely on costly wearable devices and require three to four hours to complete. In contrast, the AI Markerless Motion Capture System developed by the Software Technology Institute of the Institute for Information Industry requires only four high-resolution cameras, with no sensors or special equipment worn by the subject. Individuals can undergo gait and posture analysis in everyday clothing, completing the entire assessment in just five minutes. The system achieves accuracy rates of up to 95 percent while significantly improving diagnostic efficiency. The AI Markerless Motion Capture System enables precise evaluation of joint mobility and fall risk, providing scientific support for rehabilitation planning for patients with spinal cord injuries and degenerative diseases. The system has already been deployed at Kaohsiung Chang Gung Memorial Hospital and Taipei Veterans General Hospital.
In terms of traffic safety, the sight of vulnerable pedestrians crossing intersections often raises concern. As countdown timers tick away, wheelchair users and individuals with limited mobility must strain to cross before the light changes, while also remaining alert to turning vehicles that may overlook wheelchair-level blind spots. To address this risk, III has developed the i-RSU Intelligent Roadside Unit, which uses AI-based image recognition to detect wheelchair users crossing intersections in real time. The system then alerts nearby drivers via electronic signboards, warning of wheelchair pedestrians ahead and prompting vehicles to slow down and yield. The system received first prize in the 2024 WITSA ICT Excellence Awards – Smart Cities Awards category. The underlying technology has already been deployed at more than 200 high-risk intersections and around medical facilities across Taiwan, significantly enhancing pedestrian safety and fostering a more inclusive and age-friendly road environment.
I-Heng Meng, Director General of the Software Technology Institute at III stated, “Amid the accelerating pace of population aging, challenges related to care and safety are no longer the sole responsibility of the healthcare sector but issues for the entire social system to deal with altogether. By embedding AI into frontline settings such as long-term care, rehabilitation, and transportation, we aim not merely to provide tools, but to translate the accumulated experience and expertise of professional caregivers into capabilities that technology can inherit and transmit. Through these four smart care solutions, III enables AI to function like a seasoned mentor—making accurate judgments and delivering timely support at critical moments and becoming the most reliable partner for governments, social welfare agencies, and families on the care front. Looking ahead, the Software Technology Institute will continue to advance applications in long-term care and healthcare, working with partners to build a safer, more resilient, and more compassionate age-friendly society.”