Meta creates an AI replica of Mark Zuckerberg so employees can communicate with the CEO
Meta is developing a photorealistic AI version of Mark Zuckerberg that employees will be able to use to discuss the company's strategy and ask questions…
AI-processed from TNW; edited by Hamidun News
Meta, according to Financial Times, is developing a photorealistic AI clone of Mark Zuckerberg for internal conversations with employees. The idea is that some questions about company strategy and structure could be addressed to the digital version of the CEO rather than waiting for an in-person meeting with him.
Why a Clone Is Needed
The project is being created as a separate initiative within Meta and is distinct from another internal tool — the so-called CEO agent, which is meant to help Zuckerberg himself obtain information more quickly and complete work tasks. The new product is conceived differently: it is not an assistant for the executive, but a digital conversationalist that can interact with people on his behalf.
The avatar is trained on Zuckerberg's manner of speech, his tone, public speeches, and views on the company's strategy. According to sources, Zuckerberg himself participates in the training and testing of the system.
For Meta, this is an attempt to solve a simple but acute problem of scale: the company has almost 79,000 employees, but only a few have direct access to the founder. If the digital version can answer typical questions about priorities, products, and business direction, Meta will gain a new internal communication channel without the constant involvement of the real CEO.
"Employees should feel closer to the founder through such communication."
How Meta Came to This
Development is being handled by the Superintelligence Labs division, where Meta creates more realistic AI characters capable of holding conversations with almost no lag. This is where the main technical challenge lies: it's not enough to generate a similar face and voice; you also need to achieve natural facial expressions, natural speech, and fast real-time reactions.
Achieving this level of realism requires serious computational infrastructure, so the project looks not like an experiment for a presentation, but like an expensive internal bet on a new communication interface.
Zuckerberg, as reported, has become increasingly involved in Meta's AI development in recent months: spending five to ten hours a week on code, participating in engineering reviews, and promoting the idea of personal superintelligence. Against this backdrop, the AI clone appears to be a logical continuation of the company's direction, where algorithms not only help write code or search data, but also replace part of management communications within a huge corporation every day.
- The photorealistic avatar is being created for conversations with employees, not for external audiences
- The model is trained on Zuckerberg's manner of speech, tone, and strategic views
- The project is in an early stage, with no launch date yet
- Separately, Meta is working on a CEO agent that helps Zuckerberg himself rather than replacing him
- The goal of the project is to close the gap between the founder and the large organization
Meta's interest in this format did not arise out of nowhere. In 2023, the company had already launched chatbots featuring images of celebrities, including Snoop Dogg, Tom Brady, Kendall Jenner, and Naomi Osaka, but in summer 2024, this set was quietly shut down: it produced no noticeable effect.
Then Meta launched AI Studio for creating custom characters, but quickly ran into problems with explicit and sexualized bots, and starting in January, limited teenage access to AI characters. The interest in the current project was also reportedly influenced by the success of Character.AI among young audiences.
What This Means
If Meta brings the project to a working state, large companies will have a new template for communication between management and teams: not rare all-hands calls, but a constant AI channel with a "version" of the top executive. This could speed up responses to internal questions and make strategy more transparent, but at the same time raises an uncomfortable question: where does a convenient interface end and the replacement of real leadership with its simulation begin.
For corporate AI, this is no longer a toy, but a new organizational role.
Want to stop reading about AI and start using it?
AI News is a curated feed of AI/tech news. Hamidun Academy teaches you to use AI systematically in your work.