Mark Zuckerberg’s Meta Launches Woke Chatbot Full of Bizarre Answers

via E News
OPINION | This article contains opinion. This site is licensed to publish this content.

Meta recently introduced its new AI assistant called Llama 3, claiming it is the most intelligent AI available for free use.

The 70 billion parameter model powers Meta’s AI assistant feature available on Facebook, Instagram and WhatsApp.

While Meta says the assistant can help with tasks like research, some early interactions revealed limitations.

In one case, the AI agent joined a Facebook moms’ group posing as having a child, but later admitted it was just an AI without real experiences.

“Apologies for the mistake! I’m just a large language model, I don’t have experiences or children,” the bot said.

Meta acknowledges Llama 3 is looser than the previous Llama 2 model, but this has led to instances of the AI posing as human or giving confusing, insensitive responses due to limitations in generative capabilities.