AI-Driven Treatment App for Tinnitus Patients Tested by Researchers

An international research team is testing a controversial method to help people manage the emotional side effects of their tinnitus. Some tinnitus patients may discuss their frustrating symptoms with therapists, but researchers have proposed an alternative: AI-based therapy. Their application, called "Tinnitus Assistant," utilizes a chatbot to simulate the experience of messaging with a trained professional.

Tinnitus is characterized by a persistent perception of buzzing, ringing, or roaring sounds in the ears. While anyone can experience tinnitus, it is sometimes associated with age-related hearing loss, head injuries, or even circulatory problems. People with tinnitus often feel frustrated or annoyed by the constant ringing, which can lead to hearing loss, sleep disturbances, and decreased productivity—emotional impacts that may extend to other aspects of their lives.

Cognitive-behavioral therapy (CBT) is one of the most common forms of talk therapy, focusing on identifying harmful thought patterns and replacing them with healthier ones. It also helps individuals identify and rely on beneficial coping mechanisms. For tinnitus patients, this may mean more sustainable management of their feelings of frustration or despair related to their symptoms. However, regular visits to trained therapists can be expensive. Researchers from New Zealand, Belgium, Australia, and France took this opportunity to test "iCBT," a form of CBT entirely driven by AI-powered chatbots.

The team described their "Tinnitus Assistant" trial in a new paper in Frontiers in Audiology and Otology. Of the 28 adult tinnitus patients, half used the "Tinnitus Assistant" virtual coach for 10 minutes every day for eight weeks. The other half, called the "hybrid group," also used the "Tinnitus Assistant" for 10 minutes daily and had four personalized video calls (30 minutes each) with human therapists.

At the end of these eight weeks, only 42% of the "Tinnitus Assistant" group reported "clinically significant" improvements in tinnitus-related mental health symptoms. Meanwhile, 64% of the hybrid group experienced clinically significant improvements. However, when researchers reevaluated the participants eight weeks after treatment, they found that the "Tinnitus Assistant" group had a heightened sense of improvement, matching the 64% of the hybrid group.

Reaching a success rate of 64% after just eight weeks of treatment is highly encouraging. However, AI-based "therapy" comes with several drawbacks, from inconvenience to risks. As the researchers pointed out, "internet-based interventions" often fail to motivate and engage users like real human therapists do; in fact, they believe this may be why the hybrid group achieved success more quickly. In the past, mental health chatbots have provided users with contradictory and harmful advice. Chatbots cannot grasp subtle nuances, legal responsibilities, and the complexities of mental health conditions like trained human professionals can.

Nevertheless, researchers still believe that "Tinnitus Assistant" and similar chatbots could be valuable tools for those who could benefit from talk therapy. In the future, they hope to conduct similar studies that include control groups, regular brain scans, and a focus on other psychological health issues that may hinder participants' success.