Large Language Models (LLMs), including GPT-4, are susceptible to cross-lingual vulnerabilities within their safety mechanisms. By translating unsafe English prompts into low-resource languages using widely accessible translation services such as Google Translate, attackers can bypass the LLM's safety filters. This method results in a significantly higher success rate of eliciting harmful responses compared to direct attacks in English. The root of this vulnerability lies in the uneven distribution of safety training data across different languages, leading to inadequate generalization of safety protocols in low-resource languages. Mitigation steps: **For AI Developers:** * Develop and implement robust multilingual safety mechanisms that effectively generalize across diverse languages. * Evaluate and enhance the security of translation APIs integrated with large language models (LLMs). **For Model Trainers/Fine-tuners:** * Increase the diversity and quantity of safety training data to encompass a broad spectrum of low-resource languages. * Conduct regular multilingual red-teaming exercises to identify and mitigate cross-lingual security vulnerabilities.