In a significant move for real-time communication across different languages, Meta Platforms, the parent company of Facebook, released a cutting-edge AI model this Tuesday. This new technology, named SeamlessM4T, can translate and transcribe speech in almost 100 languages.
What Does It Do?
The SeamlessM4T model can support translations between text and speech in a wide range of languages, as well as full speech-to-speech translation for 35 languages. Before now, these functions were only available in separate models. But Meta has successfully combined them.
The Future in the Metaverse
Mark Zuckerberg, CEO of Meta, has big plans for tools like this one. He wants them to help people from all over the world talk to each other in the metaverse. The metaverse is a network of connected virtual worlds where Meta is focusing its future efforts.
Available to the Public
Meta is sharing this model with the public for non-commercial use, as mentioned in their blog post. This adds to the list of mostly free AI models that the world’s biggest social media company has released this year.
Another significant release was a large language model named Llama. This posed a serious challenge to other models from big names like Microsoft-backed OpenAI and Google’s Alphabet.
Why Is It Free?
Zuckerberg believes in an open AI system, where everyone benefits. Meta can gain more by letting many people help create tools for its social platforms, rather than charging money for them.
Legal Questions
But Meta also faces legal issues. Some of these issues relate to the data used to create its models. In July, comedian Sarah Silverman and two authors sued Meta and OpenAI. They claimed these companies used their books without permission as training data.
For the SeamlessM4T model, Meta used audio training data from 4 million hours of raw audio. The origin of this audio is unclear, as the company did not specify which repository it came from. Text data was pulled from Wikipedia and related sites.
A Step Forward
This move by Meta is a significant step in breaking down language barriers. It makes communication more accessible and opens doors for global interaction in the metaverse.
This news is based on an article from thestar.com.my.
