Meta is assembling a top-tier cohort of AI researchers and engineers aimed at refining the potent algorithms that keep users engaged on Facebook and Instagram, according to Business Insider, citing job postings and LinkedIn updates. Strengthening these recommendation systems is intrinsically linked to the company’s advertising operations, which remain its primary revenue stream.
Details
The newly formed group, dubbed MRS Research, is part of the Meta Recommendation Systems division, which oversees the content feeds across the company’s applications, Business Insider reports. MRS engineers the algorithms that dictate what users see and works closely with the firm’s advertising sector.
The unit has already secured talent from both TikTok and Amazon, the publication notes. Job descriptions indicate that the team will concentrate on ambitious, forward-looking AI projects and cutting-edge scientific investigations to advance Meta’s recommendation algorithms.
The head of MRS Research is Meta’s VP of Recommendation Research, Yan Song, who joined in November 2025, transferring from TikTok where he managed user base expansion and recommendation systems, as relayed by the outlet. Soon after, AI researcher Lihong Li from Amazon joined the team, as evidenced by his LinkedIn profile. Furthermore, the company has hired Xiaolong Wang, a former OpenAI researcher, and Fei Sha, a specialist from Google, according to LinkedIn data.
Meta declined to provide specifics regarding these AI appointments.
Meta’s AI Focus
This move by Meta signals its intent to leverage AI to bolster its core sources of income, specifically advertising, Business Insider observes. For instance, late in 2025, the company rolled out an AI model that, by its internal metrics, boosts advertising efficacy by showing consumers more contextually relevant ads. In December, Meta began utilizing data gleaned from user interactions with its AI (like dialogue content) to fine-tune targeting and deliver personalized content suggestions across its social platforms.
Advertising constitutes the vast majority of the company’s income. For the totality of 2025, its revenue reached \$201 billion, with advertising contributing \$196.2 billion. On March 30th, Morgan Stanley released a report projecting Meta’s ad revenue to climb by 28% in 2026, followed by a 21% increase in 2027.
The firm has been aggressively expanding its AI staff since the summer, when it announced the creation of Meta Superintelligence Labs (MSL), directed by former Scale AI CEO Alexander Wang. However, a Meta spokesperson clarified that the MRS division is separate from MSL.
The Status of Meta Shares
Meta’s stock advanced by 1.2% in trading on April 1st, though it remains down roughly 12% year-to-date. Analysts at Morgan Stanley suggest that investor sentiment regarding the stock has soured due to concerns over long-term strategic direction in generative AI, uncertainty surrounding return on invested capital, and macroeconomic, advertising, and regulatory headwinds, MarketScreener reports.
On March 24th, Meta shares plummeted by 8%, falling below the \$548 mark for the first time in 11 months. This sell-off wiped out \$119 billion in market capitalization, dropping the company out of the top seven most valuable US issuers. The stock dip was partly influenced by adverse outcomes in two high-profile lawsuits concerning social media’s impact on youth. A Los Angeles court ordered Meta and Google to pay \$6 million in a case related to mental health harm linked to addiction to Instagram and YouTube, while a New Mexico court fined Meta \$375 million for inadequate protection of minors against online dangers.
Morgan Stanley maintained its “buy” recommendation for Meta stock but revised its price target downward from \$825 to \$775. This revised target suggests a potential upside of 35% relative to the last closing price.
The consensus view from Wall Street regarding Meta shares remains favorable: 61 out of 69 covering analysts advise purchasing the stock, while six hold a neutral stance. There are currently no “sell” recommendations outstanding.
About the Author
