- Meta is developing a photorealistic AI avatar of Mark Zuckerberg for employee interaction.
- The project is part of a wider shift towards an AI-first workplace strategy.
- The company may extend the technology to creators and influencers if successful.
Meta is working on an AI-powered version of Mark Zuckerberg, designed to interact with employees and simulate direct communication with the chief executive, as the company accelerates its shift towards an AI-first workplace.
According to a report by the Financial Times, the social media company is developing a photorealistic 3D AI avatar that mirrors Zuckerberg’s voice, tone, appearance and communication style. The system is being trained on a mix of his public statements, behavioural patterns and recent strategic thinking, with the aim of enabling real-time conversations with staff.
The project appears to be part of Meta’s broader push into artificial intelligence, which Zuckerberg has described as a move towards “personal superintelligence”, as quoted in a news report.
A digital stand-in for leadership
The idea, as it stands, is to create a version of the CEO that can respond to questions, provide feedback and engage with employees, particularly in a large organisation where direct access to leadership is limited.
Zuckerberg is reportedly involved in testing and refining the system himself, spending time training the model to better reflect his communication style. The intention seems to be less about replacing leadership and more about scaling its presence across the company.
The AI avatar is separate from another internal tool, often referred to as a “CEO agent”, which is being developed to assist Zuckerberg with his own tasks such as retrieving information and managing workflows. That system is aimed at improving efficiency at the top level, while the avatar focuses on communication within the organisation.
The project is being led by Meta’s Superintelligence Labs, which is working on creating lifelike AI characters capable of real-time interaction. Developing such systems is resource-intensive, requiring significant computing power to generate realistic facial movement, voice output and low-latency responses.
Meta has also been investing in improving voice technology, following its acquisitions of PlayAI and WaveForms. These capabilities are expected to play a role in making the AI avatar sound more natural and responsive.
At the same time, the company is encouraging employees to integrate AI tools into their daily work. Internal initiatives include the use of agentic systems such as OpenClaw and participation in AI-focused skill assessments, which reportedly include coding and system design exercises.
Zuckerberg himself is said to be spending between five and ten hours a week working on AI-related projects and attending technical reviews, reflecting the level of focus being placed on this transition.
Getty Images
Beyond internal use
While the AI avatar remains in its early stages, Meta is understood to be exploring wider applications. If the technology proves effective, similar tools could be offered to creators and influencers, allowing them to build digital versions of themselves to interact with audiences.
This would build on Meta’s existing AI Studio platform, which already allows users to create digital personas, but with a more advanced and lifelike layer of interaction.
For now, the experiment raises broader questions about how leadership, communication and presence might evolve in large organisations as artificial intelligence becomes more embedded in everyday work.
