Domain-Invariant and Language-Model-Assisted Acoustic Side-Channel Attacks on Keyboards
Researchers demonstrate acoustic side-channel attacks that recover typed text from keyboard sounds in diverse, noisy environments. The approach utilizes language models to improve keystroke inference accuracy without device access.
why this severityResearch paper demonstrating an attack with no known exploitation.
- first report
- 11 Nov 2026
- last update
- 11 Nov 2026
Timeline · 1
Sources · 1
Research and catalogues
- 2026-11-11Domain-Invariant and Language-Model-Assisted Acoustic Side-Channel Attacks on KeyboardsarXiv · arxiv.org/abs/2611.13922
Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.
Classification
- lane
- Research
- AI role
- AI-enabled attack
- categories
- Adversarial evasion
- severity
- 1 · Research / proof of concept
- targets
- Individuals
- owasp llm top 10
- no entry for these categories
- dates
- happened 11 Nov 2026 · first reported 11 Nov 2026
- origin
- News & research feeds
- record
- confidence 0.90 · created 2026-11-11 23:41Z · updated 2026-11-11 23:41Z
Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned
Copy a brief
This record as a short Markdown brief, for a ticket, a chat or a coding agent.
the brief as text
# Domain-Invariant and Language-Model-Assisted Acoustic Side-Channel Attacks on Keyboards https://gcdradar.vercel.app/incidents/domain-invariant-and-language-model-assisted-acoustic-side-c · AIR-2026-1914 · live - Lane: Research - Severity: 1 (Research / proof of concept): Research paper demonstrating an attack with no known exploitation. - Fix status: Fix status unknown - First reported: 2026-11-11 - AI role: AI-enabled attack ## Summary Researchers demonstrate acoustic side-channel attacks that recover typed text from keyboard sounds in diverse, noisy environments. The approach utilizes language models to improve keystroke inference accuracy without device access. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-11-11: Domain-Invariant and Language-Model-Assisted Acoustic Side-Channel Attacks on Keyboards (https://arxiv.org/abs/2611.13922) ## Sources - 2026-11-11 arXiv: Domain-Invariant and Language-Model-Assisted Acoustic Side-Channel Attacks on Keyboards https://arxiv.org/abs/2611.13922 Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- liveQ-Learning Fault Injection Attacks on Q-Learning: Analysis, Vulnerabilities, and Countermeasures
Research · since 13 Nov 2026
AI - liveHijacking a Trusted LLM in Quantized Deployments via Disguised Adapters
Research · since 13 Nov 2026
AI - liveQwen Stress-testing sandbagging detection under adaptive attack
Research · since 13 Nov 2026
AI - liveFake news articles change forecasts of LLM forecasting systems
Research · since 13 Nov 2026
AI - liveChatter Attack Resource Consumption Technique for Large Language Models
Research · since 10 Nov 2026
AI - liveRAG SNAPRAG: Cost-efficient knowledge extraction attack on RAG systems via adversarial perturbations
Research · since 10 Nov 2026
AI