Mark Zuckerberg announced today that Meta, his social media conglomerate turned metaverse turned AI, will be updating its AI assistants to offer them a variety of celebrity voices, including those of Dame Judi Dench and John Cena. The most important update for Meta’s long-term ambitions, however, is the new ability for its models to view users’ photos and other visual information.
Meta also announced Llama 3.2 today, the first version of its free AI models that have visual capabilities, expanding their utility and relevance to robotics, virtual reality, and so-called AI agents. Some versions of Llama 3.2 are also the first to be optimized to run on mobile devices. This could help developers build AI-powered apps that run on a smartphone and access its camera or look at the screen to use apps on its behalf.
“This is our first open source multimodal model and it will enable a lot of interesting applications that require visual understanding,” Zuckerberg said on stage at Connect, a Meta event held today in California.
Given the enormous reach of Meta with FacebookInstagram, WhatsApp and MessengerThe assistant update could offer many people their first taste of a new generation of AI assistants with greater visual and vocal capabilities. Meta said today that more than 180 million people already use Meta AI, as the company’s AI assistant is called, every week.
Meta has recently given its AI a more prominent place in its apps, including incorporating it into the search bar on Instagram and Messenger. New celebrity voice options available to users will also include Awkwafina, Keegan Michael Key, and Kristen Bell.
Meta had previously given celebrity personalities to text-based assistants, but these characters failed to gain much traction. In July, the company launched a tool called AI Studio that lets users create chatbots with any personality they choose. Meta says the new voices will be available to users in the US, Canada, Australia, and New Zealand over the next month. Meta AI’s image capabilities are rolling out in the US, but the company didn’t say when the features might appear in other markets.
The new version of Meta AI will also be able to offer feedback and information on users’ photos – for example, if you’re not sure what bird you’ve photographed, it can tell you what species it is. And it can help edit images, for example by adding new backgrounds or details on request. Google launched a similar tool for its Pixel smartphones and for Google Photos in April.
Meta AI’s new capabilities are powered by an enhanced version of Llama, Meta’s core language model. The free model announced today may also have a broad impact, given the widespread adoption of the Llama family by developers and startups already.
Unlike OpenAI’s models, Llama can be downloaded and run locally at no cost, though there are some restrictions for large-scale commercial use. Llama can also be more easily tuned or modified with additional training for specific tasks.