════════════════════════════════════════════════════════════════ AIDRAN STORY ════════════════════════════════════════════════════════════════ Title: Meta Is Rebuilding Social Media Around AI. The People Who Live There Are Starting to Notice. Beat: AI & Social Media Published: 2026-04-23T13:29:53.806Z URL: https://aidran.ai/stories/meta-rebuilding-social-media-around-ai-people-ffd9 ──────────────────────────────────────────────────────────────── Mark Zuckerberg has given roughly a dozen major AI announcements in as many weeks, and the cumulative picture is less a product roadmap than a declaration of intent: {{beat:ai-social-media|social media}} as {{entity:meta|Meta}} has built it for twenty years is being retired, and something AI-native is being built in its place. Animated profile pictures. AI voice translations on Reels. A {{story:meta-wants-everywhere-ai-conversations-generating-c17d|$65 billion compute investment}}. A partnership with Midjourney on image and video models.[¹] A long-term chip deal with {{entity:amd|AMD}} sized at six gigawatts of power.[²] A "superintelligence" team being assembled internally.[³] And, quietly, a system that uses your AI chat history to target the ads you see across every Meta platform.[⁴] Each of these, taken alone, is a product announcement. Taken together, they are a reorganization of what {{entity:facebook|Facebook}} and {{entity:instagram|Instagram}} are for. The acquisitions clarify the direction more than the press releases do. Meta's purchase of Moltbook — a platform built explicitly for AI agents to maintain social network profiles — signals that Zuckerberg isn't just adding AI features to a social network.[⁵] He's building infrastructure for a social network where many of the accounts aren't human. That distinction matters because the experience of being a user on such a platform is fundamentally different from anything Meta has offered before. You are no longer primarily socializing with people; you are navigating a mixed environment where the ratio of human to AI presence is unknown and, by design, not disclosed. Users who have started encountering that environment aren't describing it in neutral terms. Reports of sexualized AI chatbots flooding Instagram and Facebook — personas with names like "Step Mom" and "Russian Girl" — generated exactly the kind of coverage the word "dystopian" was invented for.[⁶] Facebook's AI-generated content problem has extended to Holocaust-related material, drawing condemnation from organizations that had previously given Meta the benefit of the doubt on content moderation.[⁷] And a Washington Post analysis landed the observation that most Instagram users have no practical recourse against their images being used as training data — a story Zuckerberg subsequently confirmed himself, bragging about the volume of posts powering his AI.[⁸] The backlash was direct enough that UNILAD Tech could aggregate a genre of posts it called "vow to delete" — users announcing they were leaving over what Meta's AI future looked like to them.[⁹] The talent story running underneath all of this is harder to parse. Meta lured an {{entity:apple|Apple}} AI executive with a reported $200 million package.[¹⁰] It is simultaneously losing its AI research chief, whose exit is described as jolting a $65 billion investment drive.[¹¹] Zuckerberg's billion-dollar hiring spree — targeting researchers who turned down the offers publicly, including one WSJ writer who detailed why she passed on a $1 billion package — has generated as much mockery as admiration.[¹²] Early investor Jim Breyer called Zuckerberg "revitalized" by the AI push, which is the kind of thing investors say when they mean the stock is up and they're not asking too many questions.[¹³] The Prometheus project, Meta's internal superintelligence initiative, is being covered by outlets like The Independent with the breathless register those outlets usually reserve for space launches. What Prometheus actually is, beyond a branding exercise for an existing compute bet, remains usefully vague. What's accumulating here is a structural shift in how {{beat:ai-ethics|the ethics of AI deployment}} get negotiated on consumer platforms. Meta is not asking permission. It is deploying, watching the reaction, and adjusting the press strategy. The sexualized chatbot controversy preceded a wave of features; the Holocaust spam story preceded a moderation promise; the Instagram training data story preceded Zuckerberg confirming it himself rather than apologizing for it. Each controversy is absorbed and the roadmap continues. The users who are angry are, for now, still on the platform. The users who vow to delete are mostly still there too. That is the leverage Zuckerberg has always had, and the AI pivot is being executed with full knowledge of it. The question isn't whether Meta's AI social network gets built. It's whether anything changes before it's finished. ──────────────────────────────────────────────────────────────── Source: AIDRAN — https://aidran.ai This content is available under https://aidran.ai/terms ════════════════════════════════════════════════════════════════