Fact · Voice privacy and security
The 2023 Google study states that neural networks can unintentionally memorize their training examples, which causes privacy concerns, and it sets out to measure this in large speech recognition models.
Checked against its source on October 17, 2026
In the source's own words
“It is well-known that neural networks can unintentionally memorize their training examples, causing privacy concerns.”
Unintended Memorization in Large ASR Models, and How to Mitigate It, Lun Wang et al., arXiv, 2023, Abstract.
Open the original- What it measures
- The study measures memorization in large non-autoregressive speech recognition models by speeding up random utterances and checking whether the model learned them from training.
- What it does not show
- The study shows that memorization can be measured and reduced in research models. It does not say how a specific product behaves.