Scientists at the University of Manchester have developed and tested an AI-powered hearing test that can reliably check a person’s hearing on a computer or smartphone without clinical supervision.
According to the research, published in the journal Trends in Hearing, the high-tech test allows patients to check their hearing at home, eliminating the need for a clinical visit. By utilizing AI, the test is able to filter out background noise, which increases its accuracy.
Specifically, the researchers created an AI-based version of the traditional “Digits-in-Noise (DIN)” test, which integrates text-to-speech (TTS) and automatic speech recognition (ASR) technologies.
The result is a fully automated hearing test that can be completed in 10 minutes without the involvement of medical personnel. The test’s clinical accuracy was validated on 31 adults, including both individuals with normal hearing and those with hearing loss. This test was then compared to two conventional DIN tests. The results showed that the AI-powered test produced practically the same results as the traditional methods, confirming the technology’s reliability and accuracy.
The study’s lead author, Mohsen Fatehipar from the University of Manchester, noted that this technology signals a revolutionary change: “After testing this technology, we are confident that with the help of AI, it is entirely possible to automate a hearing test in such a way that it will be convenient to conduct it at home.”
The researchers emphasize that people with hearing problems often delay seeking help – there is an average delay of 8.9 years between needing a hearing aid and receiving one. This new system, which doesn’t require a clinic visit, will facilitate the early detection of the problem and the taking of timely action.
The scientists plan to make the program free for everyone. This decision will contribute to the development and increased accessibility of advanced speech technologies.
