Meta CEO Mark Zuckerberg announced Meta AI, the company’s new assistant powered by artificial intelligence (AI). This is his own answer to OpenAI’s ChatGPT, which integrates with Instagram, Facebook, WhatsApp, and eventually the company’s mixed reality devices.
Talk at MetaConnect event On September 27, Zuckerberg said that MetaAI is powered by Llama 2, the company’s large-scale language model, and was built in partnership with Microsoft Bing to give users access to real-time information from the internet. He explained that it was done.
“Meta AI is basically an assistant that can talk like a human.”
In addition to answering questions and conversing with users, the newly released bot also uses a new image generator called Emu, trained by Meta on 1.1 billion pieces of data, including photos and captions shared by users on Facebook and Instagram. You can also generate images using tools.
Zuckerberg pointed out the key differences between Meta AI and competitor ChatGPT, saying that rather than creating a one-size-fits-all chatbot, Meta’s overarching strategy is to create a variety of chatbots for different use cases. The aim was to create a new AI product.
As an example, he showed how Meta AI is slightly different in each of the company’s social media apps, and gave an example of how Meta AI can be added to Facebook Messenger group chats to help organize travel plans.
Zuckerberg said Meta’s chatbot isn’t just about sending useful information. It’s also designed to be conversational and fun.
Meta showcased its entertainment-focused AI products and also announced a collection of chatbots based on about 30 celebrities, including Paris Hilton, Snoop Dogg, and former NFL player Tom Brady.
According to Meta, the new AI assistant will be available on September 27 for a limited group of US-based users of Facebook Messenger, Instagram, and WhatsApp.
Meta AI will also be available to users of the company’s new smart glasses, scheduled for release on October 17th for U.S. users, and the new Quest 3 VR device.
Related: Elon Musk, Mark Zuckerberg and Sam Altman speak about AI regulation in Washington
On the same day as Meta’s Connect event, OpenAI announced that its chatbot ChatGPT will no longer be data-limited by 2021.
According to a September 27 post on X, Plus and Enterprise users with the GPT-4 model will be able to take advantage of this update immediately.
Until this update, ChatGPT suffered from ever-widening gaps in its knowledge base. Due to the nature of AI model training methods such as Generative Pre-Training Transformers (GPT), ChatGPT’s knowledge base previously ended in 2021, which is likely the year it completed its commercialization.
magazine: AI eye: Practical Uses of AI in Cryptocurrency, A Rival to Google’s GPT-4, and AI Edge for Bad Employees