# Content Anonymization for Privacy in Long-form Audio

arXiv: https://arxiv.org/abs/2510.12780 · HTML: https://mlsanity.com/paper/2510.12780/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 27 references that are in the map.
- First posted on arXiv 14 Oct 2025.

## Builds on (most cited first, 20 of 20)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [VoxCeleb2: Deep Speaker Recognition](https://mlsanity.com/paper/1806.05622/) (2018)
- [UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022](https://mlsanity.com/paper/2204.02152/) (2022)
- [Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text](https://mlsanity.com/paper/2401.12070/) (2024)
- [XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Model](https://mlsanity.com/paper/2406.04904/) (2024)
- [Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation](https://mlsanity.com/paper/2202.12233/) (2022)
- [Same Author or Just Same Topic? Towards Content-Independent Style Representations](https://mlsanity.com/paper/2204.04907/) (2022)
- [The VoicePrivacy 2024 Challenge Evaluation Plan](https://mlsanity.com/paper/2404.02677/) (2024)
- [A Deep Metric Learning Approach to Account Linking](https://mlsanity.com/paper/2105.07263/) (2021)
- [Encoder-Decoder Based Attractors for End-to-End Neural Diarization](https://mlsanity.com/paper/2106.10654/) (2021)
- [Keep It Private: Unsupervised Privatization of Online Text](https://mlsanity.com/paper/2405.10260/) (2024)
- [Can Authorship Attribution Models Distinguish Speakers in Speech Transcripts?](https://mlsanity.com/paper/2311.07564/) (2023)
- [JAMDEC: Unsupervised Authorship Obfuscation using Constrained Decoding over Small Language Models](https://mlsanity.com/paper/2402.08761/) (2024)
- [Revisiting Speech Content Privacy](https://mlsanity.com/paper/2110.06760/) (2021)
- [New Challenges for Content Privacy in Speech and Audio](https://mlsanity.com/paper/2301.08925/) (2023)
- [Use Cases for Voice Anonymization](https://mlsanity.com/paper/2508.06356/) (2025)
- [Exploratory Evaluation of Speech Content Masking](https://mlsanity.com/paper/2401.03936/) (2024)
- [SecureSpeech: Prompt-based Speaker and Content Protection](https://mlsanity.com/paper/2507.07799/) (2025)
- [The Impact of Automatic Speech Transcription on Speaker Attribution](https://mlsanity.com/paper/2507.08660/) (2025)

## Built on by (most cited first, 3 of 3)

- [Anonymization, Not Elimination: Utility-Preserved Speech Anonymization](https://mlsanity.com/paper/2604.17000/) (2026)
- [DialogPII: A multilingual dataset of synthetic dialog transcripts to detect personal information](https://mlsanity.com/paper/2606.30312/) (2026)
- [AegisVox: Factorized Privacy and Clinical Preservation for Long-Form Speech](https://mlsanity.com/paper/2610.14643/) (2026)

## Abstract (arXiv)

Voice anonymization techniques have been found to successfully obscure a speaker's acoustic identity in short, isolated utterances in benchmarks such as the VoicePrivacy Challenge. In practice, however, utterances seldom occur in isolation: long-form audio is commonplace in domains such as interviews, phone calls, and meetings. In these cases, many utterances from the same speaker are available, which pose a significantly greater privacy risk: given multiple utterances from the same speaker, an attacker could exploit an individual's vocabulary, syntax, and turns of phrase to re-identify them, even when their voice is completely disguised. To address this risk, we propose a new approach that performs a contextual rewriting of the transcripts in an ASR-TTS pipeline to eliminate speaker-specific style while preserving meaning. We present results in a long-form telephone conversation setting demonstrating the effectiveness of a content-based attack on voice-anonymized speech. Then we show how the proposed content-based anonymization methods can mitigate this risk while preserving speech utility. Overall, we find that paraphrasing is an effective defense against content-based attacks and recommend that stakeholders adopt this step to ensure anonymity in long-form audio.