Towards Probing Speech-Specific Risks in Large Multimodal Models: A Taxonomy, Benchmark, and Insights

Hao Yang, Lizhen Qu, Ehsan Shareghi, Reza Haf


Abstract
Large Multimodal Models (LMMs) have achieved great success recently, demonstrating a strong capability to understand multimodal information and to interact with human users. Despite the progress made, the challenge of detecting high-risk interactions in multimodal settings, and in particular in speech modality, remains largely unexplored. Conventional research on risk for speech modality primarily emphasises the content (e.g., what is captured as transcription). However, in speech-based interactions, paralinguistic cues in audio can significantly alter the intended meaning behind utterances. In this work, we propose a speech-specific risk taxonomy, covering 8 risk categories under hostility (malicious sarcasm and threats), malicious imitation (age, gender, ethnicity), and stereotypical biases (age, gender, ethnicity). Based on the taxonomy, we create a small-scale dataset for evaluating current LMMs capability in detecting these categories of risk. We observe even the latest models remain ineffective to detect various paralinguistic-specific risks in speech (e.g., Gemini 1.5 Pro is performing only slightly above random baseline). Warning: this paper contains biased and offensive examples.
Anthology ID:
2024.emnlp-main.614
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10957–10973
Language:
URL:
https://aclanthology.org/2024.emnlp-main.614
DOI:
Bibkey:
Cite (ACL):
Hao Yang, Lizhen Qu, Ehsan Shareghi, and Reza Haf. 2024. Towards Probing Speech-Specific Risks in Large Multimodal Models: A Taxonomy, Benchmark, and Insights. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 10957–10973, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Towards Probing Speech-Specific Risks in Large Multimodal Models: A Taxonomy, Benchmark, and Insights (Yang et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.614.pdf