With fewer people posting original content to Facebook and Instagram, that means that one of Meta’s key market advantages is degrading, because without users generating more data and insight, Meta can’t then use that information to power its ever-expanding AI systems, which could, eventually, see it lose out to other AI projects.
That’s why Reddit has become so popular with AI projects, because Reddit discussions center around questions and answers, which directly relates to AI chatbot use. xAI brings in conversational insight from X, and Meta has been able to replicate this, to some degree, via Threads.
But what Meta really needs is more day-to-day conversation, more interaction history, more insight into the questions people are asking, and the answers they prefer.
Which is seemingly behind Meta’s latest two AI partnerships.
Earlier this month, Meta acquired AI startup Limitless, who’s main product is an AI-powered pendant that’s able to record everyday conversations, and generate summaries.
Meta will no doubt look to build the Limitless pendant into its AI glasses (Limitless says that it will continue to support the pendant for another year, but will no longer sell them), with its advanced recording capacity able to help Meta gather more data from people’s everyday conversations, lessening its reliance on, say, Facebook posts for this type of input.
And late last week, Meta also announced a new partnership with ElevenLabs, which specializes in AI voice translation and conversion.
ElevenLabs says that the partnership will initially focus on localizing Reels into various languages, while it’s also looking to bring translation tools to Horizon.
Which will be another way for Meta to bring in more translated, interpreted conversational data, which it can then use to continue to build out its AI infrastructure, with direct examples of conversations, questions, answers, and more.
It seems that Meta is conscious of the possibility of losing its edge in the market if it no longer has the biggest trove of human conversations, and again, with all the engagement gains on Facebook and IG coming from Reels viewing, it makes sense for Meta to seek out more conversational inputs, in order to maximize this element.
But we might also be set to face a new series of questions about the legalities of recording people’s everyday conversations, and whether such tools should be allowed to use this as a data input.
I mean, according to the Trump Administration’s new directive to “remove barriers to United States AI leadership,” Meta will indeed be allowed to record and use such, as restricting this would impede U.S. progress.
Maybe that’s why Meta’s making the move into advanced conversational capture now, knowing that it’ll have at least a few more years, under Trump, when it will be allowed to access this data input without being challenged by privacy regulations.
Either way, it’s another interesting step for Meta’s AI ambitions, adding more data into its LLM stream.
