How Hearing-Care Professionals View Artificial Intelligence

By Melanie Hamilton-Basich. Hearing-care professionals interviewed by Melanie Hamilton-Basich are cautiously optimistic that artificial intelligence (AI) can improve patient care and clinic operations. They also insist that it should support, rather than replace, professional judgment and human care. They cite potential gains in speech understanding in noise, adaptive auditory training, testing and administration, while stressing that a clinician must check results and tailor decisions to each patient.

Hunter Gerhart, AuD, audiology director at Livingston Hearing Aid Center in Austin, Texas, said people are generally excited about what AI might do for patients, although he and many colleagues have reservations. One obstacle is confusion about what the term AI covers and how different systems work.

Different kinds of AI

Machine learning and generative AI are related but different. Machine learning uses patterns learned from data to make predictions that can support decisions. Generative systems such as ChatGPT are trained on large data sets to create new text, images or video. Most AI applications discussed in audiology use machine learning: some audiometers can automate parts of conventional pure-tone testing, while advanced hearing-aid chips can classify sound sources and adapt in real time to a wearer's listening situation. These capabilities still require appropriate professional oversight.

AI in hearing aids

Gerhart said he has fitted increasing numbers of AI-equipped hearing aids and that patients in his clinic report better hearing in noisy settings. He attributes those reports to sound-source classification that can help speech stand out from noise. Julie Prutsman, AuD, founder of Sound Relief Tinnitus and Hearing Center in Colorado, also believes AI can help separate speech from noise.

Prutsman said AI-based adjustments can help patients who otherwise feel unsure which setting to change, or by how much, in a noisy room. In her view, technology that helps someone remain confident and involved in social settings may help prevent withdrawal. She described AI controls in apps and hearing-aid chips as promising and expected them to improve further. These clinicians' observations are not comparative trial results.

Recent hearing-aid announcements

• GN ReSound announced AI hearing aids supporting Bluetooth LE Audio.
• Oticon Jet PX MiniRITE was described as using deep-neural-network technology.
• Phonak introduced its real-time AI Infinio hearing aids.
• Signia introduced its Pure Charge&Go BCT IX series.
• Starkey introduced products with advanced processing and connectivity.
• Unitron introduced the AI-powered Smile hearing system with cross-platform compatibility claims.
• Widex introduced Widex Allure hearing aids and cloud-based fitting software.

Auditory training

AI can also be used in auditory rehabilitation. Prutsman cited LACE AI Pro, an auditory-training app she says helps patients work on understanding speech in noise. Its interactive lessons increase in difficulty as a user improves and become easier when the system detects difficulty, with the aim of avoiding frustration and encouraging continued practice. Prutsman said patients need not notice every adjustment made by the algorithm to benefit from the training.

Hearing testing

Audiologists expressed more caution about AI in hearing tests. Jessica Stiel, AuD, of Ohio State University's Wexner Medical Center said AI might help screen normal hearing or typical age-related loss, but clinicians must recognize warning signs such as asymmetric or conductive hearing loss that call for referral. She compared this with an initial software-assisted hearing-aid fitting: professionals still adjust treatment in response to an individual's experience, audiogram, performance and concerns.

Prutsman saw potential in AI taking on some basic tests only under direct hearing-care-professional supervision. Accurate input data and cross-checking are essential, she said, or a system may miss or misclassify a condition. She called for clear standards. She also argued that, if these safeguards are met, AI could help clinics respond to a predicted shortage of audiologists and growing patient demand.

Gerhart worried that AI testing could go too far. He said people cannot reliably diagnose their own hearing loss and that an accurate assessment of type and severity requires a licensed professional. He did not want AI to take over diagnosis or treatment.

Clinic operations

AI is also entering administrative software through speech-to-text, automated text or email replies, multilingual translation and scheduling. Stiel said her clinic uses AI for appointments and finds it useful. Prutsman's clinic had not adopted such software at the time of the article, but she was open to tools that could reduce the time spent planning daily patient schedules.

Marketing

Prutsman said her husband, who handles their clinic's marketing, uses generative AI such as ChatGPT to draft promotional materials. In her account, it saves time on repeated polishing and allows a core message to be adapted to different formats, tones and blog titles. As the tool is used, she said, it can better reflect their preferences. She warned that a person must still review every generated piece for errors and fabricated information.

A tool, not a replacement

Stiel called AI powerful but no substitute for clinical experience. Gerhart said he sees people post hearing-test results in Facebook groups and ask professionals to interpret them; as AI becomes more accessible, he worries they may turn to an automated answer before seeking qualified advice. Stiel described a company's offer to pay audiologists to answer hearing-related questions, which she suspected might be used to train an AI answer database. She did not want that system to displace professional care.

For Stiel, hearing care is too individualized for an algorithm to replace clinical judgment or an understanding of a patient's circumstances. Gerhart likewise emphasized one-to-one interaction with a hearing-care professional as an essential part of good care that AI cannot provide.

Original article: The Hearing Review.