Impaired Prosodic Processing but Not Hearing Function Is Associated with an Age-Related Reduction in AI Speech Recognition.

The Speech Type × Age Group interaction was also significant (F 1,159 = 38.141, p = 5.3 × 10 −9 , and ω 2 = 0.107), because middle-aged to older adults categorized Wavenet speech more frequently as human speech than younger adults (t 159 = 6.434, p = 1.4 × 10 −9 , and d = 1.017) and human speech less frequently as human speech, although this was only marginally significant (t 159 = 1.962, p = 0.051, and d = 0.310).

also in 5,085 other papers