When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911723801280512 |
|---|---|
| author | Züfle, Maike Niehues, Jan |
| author_facet | Züfle, Maike Niehues, Jan |
| contents | SpeechLLMs are increasingly deployed in professional settings where domain customisation is standard practice: users supply context in prompts with sensitive information, fine-tune on proprietary recordings, or both. We identify and systematically investigate an overlooked privacy risk of such customisation: a model adapted to recognise domain-specific terminology can be nudged into transcribing a phonetically similar word from its context or training data, even when a different word is spoken, thereby leaking private information. To evaluate this risk, we construct a controlled dataset and measure leakage rates across two customisation mechanisms, prompting and fine-tuning. Both mechanisms cause measurable leakage, compounding when combined. We evaluate a prompt-level mitigation strategy and analyse the accuracy-leakage trade-off across customisation approaches, finding that fine-tuning without context prompts offers the best balance. We release our code and dataset publicly. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_28211 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR Züfle, Maike Niehues, Jan Computation and Language SpeechLLMs are increasingly deployed in professional settings where domain customisation is standard practice: users supply context in prompts with sensitive information, fine-tune on proprietary recordings, or both. We identify and systematically investigate an overlooked privacy risk of such customisation: a model adapted to recognise domain-specific terminology can be nudged into transcribing a phonetically similar word from its context or training data, even when a different word is spoken, thereby leaking private information. To evaluate this risk, we construct a controlled dataset and measure leakage rates across two customisation mechanisms, prompting and fine-tuning. Both mechanisms cause measurable leakage, compounding when combined. We evaluate a prompt-level mitigation strategy and analyse the accuracy-leakage trade-off across customisation approaches, finding that fine-tuning without context prompts offers the best balance. We release our code and dataset publicly. |
| title | When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.28211 |