While there have been significant strides in machine translation due to advances in neural network approaches, we are currently far from having a single, generic AI agent capable of handling universal translation across all languages. The challenge lies in the fact that there are approximately 7000 languages spoken globally, with the majority of people knowing just one or two of them. Even though companies like Meta have unveiled AI models that can translate up to 100 languages, there is still a long way to go before achieving true universal translation.
The development of an AI agent for universal translation would require the creation of ad hoc agents, each handling the syntax of one language, which would then map that syntax information into an agreeable universal format. While tools like Google's Translation Hub use generative AI to provide self-service document translation in 135 languages, these systems still have limitations and are not yet capable of fully understanding the meaning and sentiment behind the words being translated. It is clear that generative AI has the potential to significantly impact the translation industry, but we are not quite there yet in terms of a universal translator that can accurately translate any spoken, written, or signed language.