OpenAI's GPT-4 safety systems broken by Scots Gaelic
The safety guardrails preventing OpenAI's GPT-4 from spewing harmful text can be easily bypassed by translating prompts into uncommon languages – such as Zulu, Scots Gaelic, or Hmong. Large language models, which power today's AI chatbots, are quite happy to generate malicious source code, recipes…
#openai #scotsgaelic #hmong #brownuniversity #googletranslate #googletranslateapi #innis #bengali #thai #hebrew