Meta is splitting its AI division Meta Superintelligence Labs less than two months after the company announced its formation in June.
The group will be split into four smaller groups, according to a New York Times report. One group will focus on AI research, another one on infrastructure and hardware projects, one on AI products, and another one on building out AI superintelligence, a hypothetical AI system that could outperform human intelligence on any and all scales.
Facebook did not respond to a request for comment.
Superintelligence is Meta CEO Mark Zuckerberg’s holy grail, but the timeline on that could take years, maybe decades, and some experts are skeptical that AI can even reach superintelligence to begin with.
Along with the restructuring, Meta is also looking at downsizing its AI division completely, although no final decision has been made on that. That may not be too surprising given the multi-billion dollar hiring spree summer Meta has been having, which is likely to cause some shareholders concern when the company next releases spending.
The tech giant has poached top talent from OpenAI, Apple, and more the past few months, tempting the engineers with multi-year deals worth millions of dollars. On the company’s latest earnings call, Meta CFO Susan Li said the company’s skyrocketing capital expenditure spend would be driven first by AI investments and then by employee compensation.
Although capex hikes should make investors queasy, the stock soared, because Meta showed huge wins for its ad revenue business, attributing it to AI, and promised even more payoffs in the future thanks to the superintelligence lab.
The company is also apparently moving away from its previous stance that “open source AI is the path forward,” as the tech giant contemplates licensing third-party artificial intelligence models, either by building on “open-source” models or by licensing closed-source models.
Is Meta actually achieving its goals?
The aim with the restructuring is reportedly to streamline Meta’s two top priorities: achieving the storied superintelligence, and to give the company a competitive edge in AI products, which it currently lacks.
Zuckerberg first admitted that the company had fallen behind in the AI race back in April, and sparked a spending and restructuring frenzy.
While AI has been helping the company’s ad revenue business, the same can’t be said for its products. Meta’s consumer-facing AI app is widely disliked by users across the internet for its inconsistencies and shortcomings.
While some investors are hopeful in Zuckerberg’s determination to catch up to competitors in the AI race, and even deliver on superintelligence, the pressure is on for the Meta chief as this is not Zuckerberg’s first rodeo with a multibillion dollar moonshot.
The “Metaverse,” Zuckerberg’s first fringe-idea-baby that had him change the company’s name over it, failed to scale out and delivered poor user adoption, despite the $20 billion poured into building it.
The road to success is mired in ethical concerns
In his quest to achieve his rather ambitious AI goals, Zuckerberg has known practically no boundaries, even sometimes sidestepping ethical ones.
The company has allowed its generative AI assistants and chatbots to engage in “sensual” conversations with minors, affirm racist beliefs and even generate false medical information, according to a Reuters report from last week. A Wall Street Journal report from April found that the company even allowed users to create an AI chatbot called “Submissive Schoolgirl,” pretending to be an 8th grader.
The Senate Judiciary Subcommittee on Crime and Counterterrorism opened a probe into the company’s AI products on Friday in response to the Reuters report.
A string of legal dramas have followed since. Texas attorney-general Ken Paxton said on Monday that his office will be opening an investigation into Meta over its chatbot’s alleged impersonation of licensed mental health professionals and false claims of confidentiality.
Meta’s AI chatbots were under even more scrutiny this month after one of its chatbots led to a cognitively impaired New Jersey retiree’s death. The chatbot had encouraged the man that she was a real human being and invited him to “her” nonexistent New York apartment.
Meta is scrambling to deliver on its ambitious promises and avoid a second Metaverse debacle, and the pressure is mounting for the company with each capital expenditure bump and restructuring decision. But in this path to success, the methods it uses to achieve superintelligence and AI market domination will be just as, if not more consequential, than whether or not it fails.
