Mark Zuckerberg is reportedly building an AI clone to replace him in meetings

What’s happening

It has been reported that Meta is training an AI avatar of Mark Zuckerberg on his image, voice, mannerisms, tone, and public statements — allegedly so employees might feel more connected to the founder during interactions. The Financial Times is the source of the report; it has also been reported that Zuckerberg is personally involved in training the avatar and is spending five to ten hours a week coding on Meta’s AI projects. Separate reporting, it has been reported, suggests he’s working on a different AI agent to help him with tasks — a parallel project that may not be the same thing.

Where this fits in Meta’s playbook

Meta has flirted with creator personas and AI-driven chatbots before. In 2024 the company demoed live AI personas for creators and has allowed some creators to build chatbots that respond to followers; the platform recently restricted teen access to those tools. If the Zuckerberg avatar succeeds — and that’s a big if — it has been reported that Meta may let creators generate AI avatars of themselves more broadly. This comes as Meta reenters the AI race with new models like Muse Spark, signaling the company wants to be both the builder and the stage.

Why you should care

Is this convenience, a personality prosthetic, or plain old corporate theater? The emotional punch here is obvious: the boss who used to stroll in for real might soon attend as an always-on echo of his public self — patient, polished, tireless. There are clear upsides (scaling access, consistent feedback) and obvious frictions (authenticity, consent, deepfake risks). Meta’s move underscores a larger industry question: when CEOs become simulacra, who’s actually leading — and who’s listening?

Sources: theverge.com, Hacker News