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

At a glance:

• Meta is training an AI avatar of CEO Mark Zuckerberg using his image, voice, mannerisms and public statements.
• The AI clone is intended to interact with employees, giving the impression of direct contact with the founder.
• Success could open the door for creators on Instagram and other Meta platforms to generate their own AI avatars.

What meta is doing

Meta Platforms is reportedly developing a digital replica of its co‑founder and chief executive, Mark Zuckerberg. According to the Financial Times, the project trains the avatar on a comprehensive dataset that includes Zuckerberg’s visual likeness, vocal patterns, speaking style and the content of his public remarks. The goal, as described by sources, is to let employees “feel more connected to the founder through interactions” with the synthetic version of him.

The initiative builds on earlier experiments that Meta showcased in 2024, when the company demonstrated a live AI persona for a creator. That demo hinted at a broader strategy: allowing individuals on Instagram to generate AI versions of themselves that can reply to comments or run custom chatbots. Earlier this year Meta even began restricting teen access to those chatbot tools, underscoring the sensitivity of the technology.

How the avatar is being built

Zuckerberg himself is said to be directly involved in the training process. The Financial Times reports that he spends five to ten hours each week coding on Meta’s AI projects and participating in technical reviews. By feeding the model with his own speech recordings, interview transcripts, and internal communications, engineers aim to capture not just his voice but also his decision‑making cadence and leadership tone.

A separate effort, reported by The Wall Street Journal in March, focuses on an AI agent that assists Zuckerberg with personal productivity tasks. While that agent is distinct from the public‑facing avatar, both projects share the same underlying ambition: leveraging large‑scale language and multimodal models to extend a human executive’s presence beyond physical constraints.

Potential implications for creators and the workforce

If the Zuckerberg avatar proves successful, Meta may roll out a self‑service toolkit for creators to build their own AI doubles. The 2024 demo already allowed a handful of influencers to field comments automatically, but a polished, CEO‑level replica could raise the bar for what is expected from brand‑owner interactions on the platform.

For internal staff, the avatar could become a new channel for corporate communications, onboarding, or even performance feedback. Critics warn that such synthetic leadership could blur accountability lines and erode trust if employees cannot easily distinguish between a human and an algorithmic interlocutor. Meta’s decision to limit teen access to AI chatbots suggests the company is aware of potential misuse and is attempting to balance innovation with safety.

Looking ahead

Meta has not disclosed a timeline for a public launch, and it remains unclear whether the avatar will be deployed company‑wide or limited to pilot groups. The broader industry is watching closely, as other tech giants explore similar “digital twin” concepts for CEOs and public figures. Future regulatory scrutiny may focus on disclosure requirements—whether users must be told they are speaking with an AI replica rather than a real person.

The experiment also raises questions about the future of executive presence in an increasingly remote work environment. As AI models become more capable of mimicking human nuance, the line between authentic leadership and algorithmic representation may continue to shift, reshaping how organizations think about authority and personal branding.