After the fuss about the text machine ChatGPT, the Facebook group wants to get involved in chatbots. Meta’s language model based on artificial intelligence should support researchers in their work, wrote founder and boss Mark Zuckerberg in a Facebook post on Friday. The technology is called LLaMA.
A Meta spokesman told the financial service Bloomberg that the system is not currently used in meta products such as Facebook or Instagram. It should be made available to researchers specializing in artificial intelligence.
Chatbots like ChatGPT have been trained on massive amounts of text to imitate human speech. They’re basically guessing what word to come next. In this way they form sentences and texts that could also come from a human being. At the same time, ChatGPT tests showed that the program sometimes answers questions with incorrect information and can become abusive in dialogue with users.
ChatGPT was developed by the start-up Open AI and released at the end of last year. This first encouraged Google and now Meta to share more of their own language model development over the years with the public. Meanwhile, Microsoft entered into a multi-billion dollar pact with Open AI.