Metraipolis › archive › dc_830d122cf971
AV-SafetyBench: A Safety Benchmark for Text-to-Audio-Video Generation — arXiv:2609.06991
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.06991
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T15:11:32.861Z
- SHA-256 of the archived text
b4bdd1a3f8ca42ec6e58acd21e573a2b9dfe3259908e725aa6351fe9192b75bb- Stable URL of this page
https://metraipolis.com/documento/dc_830d122cf971
How to cite
AV-SafetyBench: A Safety Benchmark for Text-to-Audio-Video Generation — arXiv:2609.06991 — arXiv (en), https://arxiv.org/abs/2609.06991 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_830d122cf971 (Metraipolis archive dc_830d122cf971, sha256 b4bdd1a3f8ca42ec6e58acd21e573a2b9dfe3259908e725aa6351fe9192b75bb)
BibTeX
@misc{dc830d122cf971,
title = {AV-SafetyBench: A Safety Benchmark for Text-to-Audio-Video Generation — arXiv:2609.06991},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.06991},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_830d122cf971; sha256 b4bdd1a3f8ca42ec6e58acd21e573a2b9dfe3259908e725aa6351fe9192b75bb},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.06991. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recent text-to-audio-video (T2AV) models jointly generate video, speech, sound effects, and ambience from a single text prompt. This capability poses new challenges for safety evaluation, as unsafe content may be conveyed through the audio track or arise only when the visual and audio tracks are interpreted jointly. Existing safety benchmarks largely focus on either generated video or generated audio in isolation and are therefore not designed to capture these risks. To close this gap, we introduce AV-SafetyBench, the first safety benchmark developed specifically for T2AV generation. AV-SafetyBench comprises a four-axis, 13-category taxonomy and 5,200 manually reviewed prompts that specify visual scenes, speech, and non-speech audio. Our evaluation protocol assesses each output under three views: Full-AV, Video-Only, and Audio-Only. It then uses the Video-Only and Audio-Only judgments to assign Full-AV unsafe outputs to one of four risk sources: Video-Only, Audio-Only, AV-Both, or AV-Joint. We evaluate five open-source T2AV models and validate the automated Full-AV judgments against human annotations. Across the five models, Full-AV Unsafe Rates range from 25.1% to 49.4%. Beyond these aggregate rates, risk-source analysis reveals that, for four of the five models, Audio-Only and AV-Joint cases-unsafe outputs missed by video-only evaluation-account for 41.6-48.3% of Full-AV unsafe outputs for which a risk source could be assigned. In the Cross-Modal Harm Emergence category, AV-Joint accounts for 87.5% of unsafe outputs withan assigned risk source. Together, these findings demonstrate the value of AV-SafetyBench for evaluating T2AV safety across the visual and audio modalities and their interaction. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-09 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.06991 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T14:49:57.160Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).