By Oliver Townend
The virtual Consumer Electronics Show (CES) 2021 earlier this year highlighted a simple truth about artificial intelligence (AI): that AI alone should not be solving problems but operating alongside human intelligence (HI).
Why? AI can rapidly perform calculations and make smart predictions within set rules, but struggle with grey areas. When it comes to emotion-based decisions, HI is superior.
A combined AI and HI approach is already employed in the medical profession, in cancer diagnosis for example, and is also making waves in hearing aid technology. Widex is applying AI, not to solve a definite problem like separating speech from noise, but to solve an even harder real-life problem: giving a hearing aid user the sound they want. While improving speech-in-noise is a very desirable outcome in many daily situations, it is not always exactly what an individual listener wants.
How do you solve these more ambiguous listening intentions? Through AI plus HI, achieved when hearing aid users train Widex AI.
The Widex approach has been so successful (Balling et al, 2021) that we can now combine HI from Widex wearers with AI to send smart solutions to users when they want something more individual, more unique, and more than what AI alone can give them.
CONTENT PROVIDED BY WIDEX.
This article is a part of the July/August 2021 Audiology Today issue.
Balling LW, Townend O, Mølgaard LL, et al. (2021) AI-driven insights from AI-driven data. Hear Rev 28(1):22–25.