Meta confirmed to TechCrunch that it’s testing Meta AI, its large language model-based chatbot, with WhatsApp, Instagram and Messenger users in India and parts of Africa. The move signals that Meta plans to tap right into a huge user base across its various apps to scale its AI offerings.
The social media giant has been moving to introduce more AI services within the wake of major AI pushes by other major tech corporations, OpenAI and others.
Meta announced plans to develop and experiment with chatbots and other AI tools in February 2023. India, where users have recently noticed the looks of the Meta AI chatbot, is a vital marketplace for the corporate: it’s home to greater than 500 million Facebook users and WhatsApp users, making it Meta's largest single market .
Users in Africa are also reporting signs of meta-AI emerging in WhatsApp.
Meta confirmed the move in a press release. “Our generative AI-powered experiences are in various stages of development and we’re publicly testing a lot of them on a limited scale,” a Meta spokesperson told TechCrunch.
Meta introduced its general-purpose assistant Meta AI in September 2023. The AI chatbot is designed to reply user queries directly in chats and offer them the flexibility to generate photorealistic images from text input. In the case of Instagram, there’s evidence that it’s also used for search queries.
Meta is a bit late in developing and introducing AI tools for its users. In some cases, the teams assumed that generative AI technology was not quite ready for prime time. OpenAI has clearly proven this unsuitable and pushed MetaAI into the background.
“The availability of ChatGPT has by some means captured the eye and excitement of the general public,” said Yann LeCun, Turing Award winner and Meta’s chief AI scientist, speaking at an “AI Day” hosted by the corporate earlier this week organized in its offices in London. “What surprised people like me about ChatGPT wasn't the technology or the performance of the system. It was the nice interest that it aroused amongst the general public. That surprised everyone. It also caught OpenAI by surprise.” Meta, he explained, felt that AI chatbots “weren’t particularly welcome” attributable to their very own efforts to bring them to market… in truth, a few of them were destroyed by humans.” Now described he describes the corporate and the broader tech community as “more open and cozy with releasing models.”
And that's exactly what Meta is doing now. More pragmatically, nevertheless, there are three the explanation why Meta could advance its AI strategy.
First, on user retention (users now expect to see and need to make use of AI tools of their apps; if Meta doesn't give them this, there's a priority that those users will churn).
Second, for investor retention (investors obviously want big returns, but within the tech space additionally they wish to see signs that Meta is committed to supporting and constructing what many imagine will probably be the following generation of computing).
Thirdly, out of its own pride (the corporate has set the pace in lots of areas resembling mobile apps, social media and promoting during the last decade and has outsized talent on its bench, including acclaimed AI academic Yann LeCun. Is that actually the case? I’ll jump the shark and miss all this?!).
Instagram and WhatsApp's huge global user base with billions of monthly energetic users actually provides Meta with a singular opportunity to scale its AI offerings. By integrating Meta AI into WhatsApp and Instagram, Facebook parent company can bring its advanced language model and image generation capabilities to a large audience, potentially eclipsing the reach of its competitors – at the least on paper.
The company individually confirmed earlier this week that it can launch Llama 3, the following version of its open source large language model, next month.
