Beyond Weaponization: NLP Security for Medium and Lower-Resourced Languages in Their Own Right
This addresses security risks for communities using lower-resourced languages, but it is incremental as it extends existing attacks to more languages.
The paper tackles the problem of language model security for lower- and medium-resourced languages by extending adversarial attacks to up to 70 languages, finding that monolingual models are often too small for sound security and multilinguality does not always guarantee improved security.
Despite mounting evidence that multilinguality can be easily weaponized against language models (LMs), works across NLP Security remain overwhelmingly English-centric. In terms of securing LMs, the NLP norm of "English first" collides with standard procedure in cybersecurity, whereby practitioners are expected to anticipate and prepare for worst-case outcomes. To mitigate worst-case outcomes in NLP Security, researchers must be willing to engage with the weakest links in LM security: lower-resourced languages. Accordingly, this work examines the security of LMs for lower- and medium-resourced languages. We extend existing adversarial attacks for up to 70 languages to evaluate the security of monolingual and multilingual LMs for these languages. Through our analysis, we find that monolingual models are often too small in total number of parameters to ensure sound security, and that while multilinguality is helpful, it does not always guarantee improved security either. Ultimately, these findings highlight important considerations for more secure deployment of LMs, for communities of lower-resourced languages.