Widex says its new Allure AI RIC hearing aids with Clarity Boost are the first devices with third‑generation audio‑specific AI, boosting audio processing to unprecedented levels.
Australian hearing professionals will gain access to the Allure AI RIC R D on 15 June 2026 after it launches at the inaugural WSA Innovation Expo in Sydney on 10 June and Melbourne on 12 June 2026.
Below, Dr FEDERICA BIANCHI explores the science, design decisions, and real‑world impact of the device.
Everyday hearing should feel natural. That’s the simple but ambitious idea that Widex has built its new Widex Allure AI RIC (receiver-in-canal) R D hearing aid around.
The device features a “breakthrough” approach to AI which, according to the manufacturer, is designed to move artificial intelligence (AI) in hearing care beyond incremental improvements and toward a more natural listening experience.
Advanced AI support is still available exactly when it is needed most, in noisy environments. The innovations mean that users can experience natural sound for daily listening, paired with on‑demand clarity in noise through Widex’s new Clarity Boost program which is powered by a dedicated AI co-processor.
The manufacturer says it’s the first hearing aid with “third‑generation audio‑specific AI”. But what does that mean, and why is it such a breakthrough?
Dr Federica Bianchi, WSA’s senior commercial audiology program manager based in Denmark, tells Hearing Practitioner Australia this represents a new generation of AI, and a step change for the industry.
“Third‑generation audio‑specific AI means that we are using an AI architecture that is purpose‑built for sound, rather than borrowed from technologies originally designed for images, to process audio end-to-end,” she explains.
“Earlier approaches either used AI to control traditional signal processing or relied on convolutional neural networks, which excel at images but struggle with the temporal nature of sound.”
Network for audio
Dr Bianchi, a researcher with a PhD in hearing science who also has a Master of Science in engineering acoustics, says that with Allure AI RIC, Widex introduced a linear recurrent neural network architecture designed specifically for audio, which is placed directly on the signal path to process audio end-to-end.
“By processing sounds sequentially while keeping a memory of past audio, this architecture can naturally capture the temporal structure of sounds like speech as it unfolds, or rhythm in music,” she says.
“The result is very effective noise reduction with dramatically better efficiency, which translates directly into reduced power consumption, hearing-aid size, and processing delay.”
While many manufacturers talk about “more powerful AI”, Widex focuses strongly on efficiency rather than brute computational power, she says.
“In hearing aids, efficiency is power. Larger models typically demand more memory, more battery consumption, and longer processing delays, each of which degrades the user experience,” she adds.
But Widex Allure AI RIC’s audio‑specific AI delivers comparable or superior denoising performance using roughly 10 times fewer parameters than typical second‑generation model, she says.
“This efficiency enables a compact design, long battery life, and critically, processing delays that stay below perceptually disruptive thresholds,” Dr Bianchi says.
Highest output SNR
According to Widex, its Allure AI RIC achieved the highest average output SNR compared to all premium competitors. But what does that mean for users in the real-world?
“Output SNR (signal‑to‑noise ratio) is a direct measure of how clearly speech emerges from background noise,” Dr Bianchi explains. “In rigorous technical studies across six realistic sound environments, Clarity Boost delivered the highest average output SNR of all four premium competitors.1
“For users, this means speech that is consistently cleaner from noise, even in restaurants, parties, train stations, and other challenging environments. Importantly, these results were shown at realistic input SNRs that mirror everyday listening conditions. This is where Widex Allure AI RIC performance truly matters.”
Avoids processed sound
In hearing aids, noise reduction often comes with complaints of “unnatural” or “processed” sound, Dr Bianchi says. However, Widex avoids that trade‑off.
“This is where Widex Allure AI RIC truly differentiates itself,” she says. “Instead of training the AI model in Clarity Boost solely to remove noise energy, we trained it using a Large Audio Foundation Model that captures those sound features that are important for human perception.
“This perceptually guided training teaches the AI to preserve qualities like naturalness, speech integrity, spatial awareness, and room acoustics, rather than aggressively stripping them away.
“We intentionally keep a controlled amount of background sound so users remain oriented and immersed in their environment.
“The result is not just clearer speech, but sound that still feels natural.”
Lowest average processing delay
Processing delay is a growing concern with AI‑based hearing aids, Dr Bianchi says, and it’s a factor addressed in the Widex Allure AI RIC design.
“Delay is critical,” she adds. “Processing delays above 10 milliseconds can negatively impact sound quality.
“Despite using a dedicated AI co‑processor, Clarity Boost does not exceed the critical 10-ms threshold by delivering the lowest average processing delay among hearing aids with an AI co‑processor.
“This is achieved through the efficiency of the third‑generation AI architecture and seamless integration with the Widex W1 chip.” The W1 chip is Widex’s fastest chip, being four times faster than its previous platform.
Summing up why Widex believes its Allure AI RIC represents the leading AI solution in the industry today, Dr Bianchi says: “Widex Allure AI RIC with Clarity Boost is the first to use an audio-specific AI architecture directly on the signal path.
“This means that the AI model in Clarity Boost operates directly and efficiently on the audio signal, which offers powerful denoising capabilities.”
She says that, with great power, comes great responsibility.
“This is why the audio-specific AI architecture is combined with perceptually motivated training, to achieve superior denoising while preserving those sound features that are important for human perception, like natural sound, signal quality, and environmental awareness,” she says.
“The result is that Clarity Boost outperformed the other manufacturers by achieving the highest average output SNR of all four competitors with AI-based denoising and the lowest average processing delay among competitors with a dedicated AI co-processor.”
Clarity Boost unlocks the full potential of AI to boost audio processing to unprecedented levels, whenever the user may need it.
The Widex Allure AI RIC hearing aid with Clarity Boost will launch at the inaugural WSA Innovation Expo in Sydney on 10 June and Melbourne on 12 June 2026. The expo is free for all audiologists and audiometrists in Australia.
Read more about the expo and register at eu.eventscloud.com/wsaexpo26.
Reference: 1. Bianchi, F, Oestergaard, M, Nielsen, JB (2026). INTRODUCING 3RD GENERATION AUDIO-SPECIFIC AI: EXPRESSIVE. EFFICIENT. ENGINEERED FOR OPTIMAL BALANCE. WidexPress – Issue #60.
