Meta is reportedly reconfiguring its AI efforts, the fourth such restructuring in six months.
The tech giant’s new artificial intelligence (AI) unit, Meta Superintelligence Labs, is expected to be divided into four groups, The Information reported Saturday (Aug. 16), citing sources familiar with the matter.
Those sources say the four groups include a new lab known as “TBD Lab,” as in “to be determined,” as well as a team focused on products such as the Meta AI Assistant. There is also a team centered around infrastructure and Meta’s Fundamental AI Research lab, which focuses on longer-term research.
PYMNTS has contacted Meta for comment but has not yet gotten a reply.
The report noted that these changes come amid a turbulent time for Meta’s AI efforts, with the company recently spending billions of dollars to recruit former Scale AI CEO Alexandr Wang and former GitHub CEO Nat Friedman to co-head Meta Superintelligence Labs.
At the same time, Meta has been on a hiring spree, onboarding dozens of experts from competitors such as OpenAI, Anthropic and Google.
Sources tell The Information that veterans of those tech giants will be involved in the new four-quadrant AI project. For example, Jack Rae, who joined Meta from Google, is expected to oversee pretraining, in which AI models learn to predict text from trillions of words gleaned from the internet and other sources.
In related news, Meta recently faced criticism over the privacy practices surrounding its AI assistant. A report earlier this month from CPO Magazine said that the tool may publicly share user prompts, and its apps may have exploited a technical loophole that lets it track Android users without their knowledge.
Consumers already have privacy-related worries surrounding generative AI. Research from the PYMNTS Intelligence report “Generation AI: Why Gen Z Bets Big and Boomers Hold Back” shows that 36% of generative AI users are nervous about these platforms sharing or misusing their personal information, while 33% of non-users have avoided adopting the technology due to the same hesitations.
Meanwhile, PYMNTS wrote last week about the hesitancy among some businesses to adopt AI, a reluctance often driven by cost concerns. Research by PYMNTS Intelligence shows that 46.7% of firms point to adoption costs as a chief concern.
The cost of the models has dropped since 2022, but the overall cost of ownership “has been resistant to declines,” said Muath Juady, founder of SearchQ.AI. “The real expenses lie in the hidden infrastructure, including data engineering teams, security compliance, constant model monitoring, and integration architects necessary to connect AI with existing systems.”