SK Telecom said Monday it has developed an artificial intelligence program that can detect dementia by listening to patients' speech, jointly with the Seoul National University College of Medicine, and begun its commercial environment test.

The AI program listens to the difference in dementia patients' voices, analyzes the voices' characteristics, and diagnoses dementia.

SKT said it is developing the program as an online application so that patients can easily use the program for testing dementia remotely without visiting medical institutions, reducing diagnostic costs.

Early detection of dementia is significant as various cognitive functions, including memory, language, and judgment, decline over the years before daily life deterioration occurs. Detecting dementia early can help patients maintain cognitive function longer by establishing health measures and preventing accompanying disorders.

The number of dementia patients and social cost is increasing every year.

The number of Korean dementia patients is expected to increase from 788,000 in 2019 to 1.36 million in 2030, with the cost of dementia treatment likely to swell to 33.7 trillion won ($29.7 billion)

The company plans to improve the accuracy of screening for dementia by testing the program at general hospitals and dementia centers with specialists and clinical psychologists of the SNU College of Medicine’s research team.

SKT will also introduce an advanced AI dementia program that uses additional information, including facial recognition, heart rate, blood pressure, and linguistic of dementia patients, such as grammar composition and language repetition, of dementia patients in 2021.

"It is a big step in the AI healthcare field to develop technology with a medical school to verify programs in a commercial environment," SKT CTO Kim Yoon said.

SNU Professor Lee Jun-young said, "Although elders who have dementia show clear differences in words and voices, it has not been easy to quantify and use them for diagnosis. Testing dementia based on voices will contribute to the early diagnosis."