Over the past yr, Meta has made its AI assistant so ubiquitous in its apps it’s virtually onerous to imagine that Meta AI is simply a yr previous. However, one yr after its launch on the final Join, the corporate is infusing Meta AI with a load of latest options within the hopes that extra folks will discover its assistant helpful.
One of many greatest modifications is that customers will be capable to have voice chats with Meta AI. Up until now, the one option to converse with Meta AI was by way of the Ray-Ban Meta good glasses. And like final yr’s Meta AI launch, the corporate tapped a gaggle of celebrities for the change.
Meta AI will be capable to tackle the voices of Awkwafina, Dame Judi Dench, John Cena, Keegan Michael Key and Kristen Bell, along with a handful of extra generic voices. Whereas the corporate is hoping the celebrities will promote customers on Meta AI’s new skills, it’s value noting that the corporate quietly phased out its movie star chatbot personas that launched finally yr’s Join.
Along with voice chat help, Meta AI can be getting new picture capabilities. Meta AI will be capable to reply to requests to vary and edit pictures from textual content chats inside Instagram, Messenger and WhatsApp. The corporate says that customers can ask the AI so as to add or take away objects or to vary components of a picture, like swapping a background or clothes merchandise.
The brand new skills arrive alongside the corporate’s newest Llama 3.2 mannequin. The brand new iteration, which comes barely two months after the Llama 3.1 launch, is the primary to have imaginative and prescient capabilities and may “bridge the gap between vision and language by extracting details from an image, understanding the scene, and then crafting a sentence or two that could be used as an image caption to help tell the story.” Llama 3.2 is “competitive” on “image recognition and a range of visual understanding tasks” in contrast with related choices from ChatGPT and Claude, Meta says.
The social community is testing different, probably controversial, methods to convey AI into the core options of its fundamental apps. The corporate will take a look at AI-generated translation options for Reels with “automatic dubbing and lip syncing.” Based on Meta, that “will simulate the speaker’s voice in another language and sync their lips to match.” It can arrive first to “some creators’ videos” in English and Spanish within the US and Latin America, although the corporate hasn’t shared particulars on rollout timing.
Meta additionally plans to experiment with AI-generated content material instantly in the primary feeds on Fb and Instagram. With the take a look at, Meta AI will floor AI-generated photos that are supposed to be personalised to every customers’ pursuits and previous exercise. For instance, Meta AI might floor a picture “imagined for you” that options your face.