While a large number of teenagers today are turning to artificial intelligence as a convenient tool to assist with academic assignments or to enhance their social interactions, thirteen-year-old Kevin Tang has taken a profoundly different approach. Rather than focusing on personal benefit, he has harnessed the power of AI as a means to improve the lives of others and to address a serious issue he encountered within his own family. Tang recounted a heartbreaking incident to Business Insider: several years earlier, his grandmother tragically slipped and fell in the kitchen, and no one realized the accident had occurred right away. By the time family members discovered her and managed to call for medical help, it was already too late to prevent catastrophic and irreversible brain damage. The experience left an enduring mark on Tang, serving as an emotional catalyst for his later innovations.
Not long afterward, he discovered that a close friend had endured a comparably painful experience—a grandparent who had also fallen but went unnoticed until the next day because the family resided in an entirely different state. These consecutive stories of preventable suffering awakened in Tang a deep sense of responsibility and purpose. He felt compelled not merely to safeguard his own relatives but to find a scalable way to aid the millions of elderly individuals who fall each year, often with life-altering consequences. This commitment set the foundation for his groundbreaking project, FallGuard.
Tang’s invention would ultimately capture the attention of the judges at the 2025 3M Young Scientist Challenge, a prestigious competition dedicated to celebrating youth-led innovation. His project won first prize, earning him a $25,000 cash award. Demonstrating foresight and maturity, Tang decided not to treat the funds as personal earnings; instead, he reinvested a substantial portion into advancing and refining FallGuard’s capabilities so that it could better serve those most vulnerable to falls.
The conception and construction of FallGuard began in the summer of 2024, when Tang immersed himself in coding and experimenting with artificial intelligence models. Over time, what started as a simple idea evolved into a sophisticated fall detection system capable of monitoring a person’s movements in real time. When the system detects an abnormal motion consistent with a fall, it immediately notifies the user’s designated family members or caregivers through the FallGuard mobile application. Impressively, the device can also recognize when someone has been motionless or lying prone for an unusually long period — a feature that could signal potential medical emergencies.
One of the notable advantages of Tang’s device is its affordability and accessibility. FallGuard operates independently of cellular service plans, eliminating messaging or subscription fees. A single device can simultaneously connect to several phones, ensuring that multiple caregivers or family members are instantly alerted if a fall occurs. This design not only makes it efficient but also adaptable for different household or caregiving arrangements.
Unlike traditional wearable fall-detection technologies that rely on users remembering to charge or wear them consistently, FallGuard functions entirely through a camera connected to a computer. As Tang has explained, once the camera is properly positioned—typically mounted on a wall—it continuously operates without the need for user interaction. Importantly, to protect user privacy, the system neither records nor transmits video footage; it processes visual data locally to minimize digital risk.
Tang acknowledges that, while the current version of FallGuard is highly promising, it has certain practical constraints. The monitored person must fall within the camera’s visible area, and each computer running FallGuard can presently support only a single camera. To overcome these limitations, Tang is developing an updated framework that would allow multiple cameras to be connected to one system, enabling full coverage of a home without the necessity of procuring multiple computers.
The technological foundation of FallGuard rests on MediaPipe, an artificial intelligence library developed by Google that allows computers to map human body movements through key points detected on the screen. Tang constructed a two-stage fall detection algorithm that analyzes a person’s posture and movement dynamics over time. Through techniques commonly used in computer vision, such as bounding boxes—which track the alteration of a person’s shape from upright to horizontal—FallGuard can determine whether a fall has likely occurred. To distinguish intentional movements from accidents, if the AI detects that an individual remains lying down, it retrospectively checks the preceding second to identify any sudden velocity decrease suggestive of a fall.
Tang continues to refine the model’s reliability and accuracy, addressing any remaining technical inconsistencies to ensure that FallGuard performs flawlessly under a wide range of conditions. During the 3M Young Scientist Challenge, he was partnered with Mark Gilbertson, a robotics and artificial intelligence specialist from 3M, who offered mentorship and practical insights. Although Tang executed all the programming and system design himself, Gilbertson contributed guidance on hardware considerations such as mounting methods and material selection. Gilbertson remarked that what immediately distinguished Tang’s work was the emotional origin of the idea—a rare combination of technical sophistication and heartfelt motivation.
Following Tang’s victory, news of FallGuard’s potential spread widely, drawing the attention of hundreds of interested families. One particularly memorable inquiry came from a man who was the primary caregiver for his wife yet was unable to hear her if she fell because of his hearing impairment. He expressed that Tang’s invention could dramatically enhance their quality of life, granting him peace of mind and ensuring his wife’s safety.
Having received recognition and financial support, Tang invested part of his award money into upgrading his equipment, including the purchase of a MacBook to code the FallGuard desktop application. This version allows ordinary users to convert their own computers into functional FallGuard devices, expanding access to those who might not otherwise obtain specialized hardware.
When reflecting upon his journey and achievements, Tang did not focus on the acclaim, the financial reward, or the prestige associated with the title of young scientist champion. Instead, he pointed with quiet pride to the actual device that now hangs on his wall—an embodiment of perseverance and compassion. He recalled how FallGuard evolved, phase by phase, from an initial experiment involving nothing more than a camera on a tripod, into a compact, wall-mounted apparatus complete with an accessible mobile application. Each iteration represented countless hours of experimentation and improvement until, at last, he reached what he considered a finished product. Tang’s story stands as a testament to how empathy fused with innovation can generate tools that genuinely enhance human welfare.
Sourse: https://www.businessinsider.com/13-year-old-won-25000-for-ai-fall-detection-device-2025-12
