Electronic stethoscopes have advanced considerably beyond traditional models. The devices can amplify respiratory sounds and use AI to flag abnormal patterns. Now, researchers at Seoul National University of Science and Technology in South Korea have developed an AI system that could help smartphones detect abnormal lung sounds without specialized equipment.

The benefit is twofold. Cost is one advantage. Electronic stethoscopes can cost hundreds of dollars, while smartphones are already in the pockets of billions of people worldwide. The other advantage is the possibility of home-based respiratory screening, potentially allowing parents to check their children’s breathing without making repeated trips to a doctor’s office.

“It could help parents home screen children between visits while keeping clinicians in the loop,” said Associate Professor Seong-Eun Kim, one of the lead researchers. “Moreover, such technology could improve remote clinic consultations and aid community healthcare workers in frontline triage in remote areas.”

The development could be an important advance in respiratory health care, particularly for young children, who are especially vulnerable to infections that can quickly become serious. By combining a smartphone’s built-in microphone with AI, researchers are exploring whether these everyday devices can serve as accessible medical screening tools.

The research, published July 30, 2026, in the IEEE Journal of Biomedical and Health Informatics, focused on one of the challenges of using smartphones for medical purposes: Their microphones were not designed to capture the subtle sounds that physicians rely on to evaluate lung function.

Traditional stethoscopes allow clinicians to listen for wheezing, crackling and other abnormal respiratory sounds that can indicate underlying health problems. Electronic versions improve upon those capabilities by amplifying sounds, reducing background noise and, in some models, using AI to help find abnormalities.

Smartphones present a different challenge. Their microphones are designed primarily for speech and general audio recording, and the quality of respiratory recordings can vary depending on the device, surrounding noise and how the phone is positioned.

To overcome those limitations, the South Korean researchers developed a deep learning system designed to recognize abnormal lung sounds despite differences between smartphone recordings and those captured by specialized medical equipment.

The study involved 611 children ages 6 and younger, a population particularly susceptible to respiratory illnesses. Of those children, 147 provided paired recordings using both electronic stethoscopes and smartphones, allowing researchers to compare sounds captured by the two devices.

They also evaluated the system using an external group of 48 children whose respiratory sounds were recorded with smartphones.

The technology relies on a technique known as domain-adaptive deep learning, which helps AI systems recognize important patterns even when the data they meet differ from the information used during training.

The researchers incorporated techniques including MixStyle, which helps models adapt to variations in recording conditions, and an Audio Spectrogram Transformer, which analyzes the frequency patterns of recorded sounds.

In simpler terms, the system learns to distinguish potentially significant respiratory sounds from the variations and imperfections introduced by different recording devices.

The results were promising.

The AI model achieved an area under the receiver operating characteristic curve (AUROC) score of 0.911 in its primary smartphone evaluation and 0.862 when tested on a separate group of smartphone recordings. These scores measure how effectively the system distinguishes normal from abnormal lung sounds, with higher values indicating better performance.

The findings do not mean smartphones can replace conventional stethoscopes or physicians, but they suggest that AI could help bridge the performance gap between consumer electronics and specialized medical equipment.

According to UNICEF, pneumonia is still the leading infectious cause of death among children younger than 5 worldwide. Many of those deaths occur in communities where prompt diagnosis and treatment are difficult to obtain.

In the United States, respiratory illnesses also place substantial demands on the health care system. The Centers for Disease Control and Prevention tracks the burden of respiratory infections, including influenza, respiratory syncytial virus and COVID-19.

A reliable smartphone-based screening system could eventually help parents and caregivers recognize potentially concerning respiratory symptoms and decide when more medical evaluation is necessary.

The technology could also strengthen telemedicine. Rather than relying solely on a patient’s description of symptoms, clinicians might someday be able to evaluate respiratory recordings collected remotely using a smartphone.

Such applications would require further development and clinical validation. The South Korean study focused on young children, and its external smartphone evaluation involved a small group. Researchers would need to determine how consistently the system performs across different ages, medical conditions, smartphone models, and real-world environments.

The research reflects a broader transformation in the relationship between smartphones and health care.

Devices once used primarily to make calls now allow people to track physical activity, check certain health indicators, access medical records, and communicate with physicians through platforms such as MyChart. AI is opening the door to more sophisticated applications, including the analysis of biological signals that traditionally needed specialized medical instruments.