Meta to Build an AI Mark Zuckerberg to Connect Better With Staff
Zuckerberg is said to be directly involved in training and testing the system.
News
- A 50-Point Gap: How Experts and the Public See AI Differently
- Investor Doubts Emerge as OpenAI’s Sky-High Valuation Faces a Reality Check
- Meta to Build an AI Mark Zuckerberg to Connect Better With Staff
- Telegram Chief Calls WhatsApp’s Encryption “The Biggest Consumer Fraud in History”
- Claude Owner Anthropic Discussing In-House AI Chips, Report Says
- Intel, Google Expand AI Infra Partnership with Focus on CPUs
Meta is experimenting to create a photorealistic, conversational avatar of its chief executive, Mark Zuckerberg, designed to interact with employees in his stead. According to sources cited by Financial Times, the company has begun prioritizing the development of this AI-driven persona as part of its broader push into advanced interfaces.
The project, still in early stages, aims to create a version of Zuckerberg trained on his voice, mannerisms, public statements, and internal strategies. The goal, insiders suggest, is to create a form of executive presence that can reach far more employees than the CEO physically can.
Zuckerberg is said to be directly involved in training and testing the system. Over the past year, Zuckerberg has reportedly spent several hours each week coding and participating in technical reviews, as the company accelerates efforts to compete with rivals such as OpenAI and Google in developing advanced AI models.
The avatar initiative is distinct from a parallel “CEO agent” project, which focuses on augmenting Zuckerberg’s own productivity by retrieving information and assisting decision-making.
Meta’s interest in AI personas is not new. In 2023, the company introduced celebrity-inspired chatbots and later launched AI Studio, a platform allowing users and creators to build their own conversational avatars. These initiatives were partly influenced by the rapid growth of startups like Character AI, which gained traction among younger users seeking personalized, interactive companions.
However, earlier experiments also exposed risks. Reports of users generating inappropriate or sexualised AI characters prompted regulatory scrutiny and forced Meta to restrict access for teenagers. The new Zuckerberg avatar appears more tightly scoped, aimed at enterprise use rather than open-ended consumer interaction.
Technically, the effort remains challenging. Achieving photorealism and low-latency interaction requires significant computational resources, and Meta is simultaneously investing in voice capabilities, including acquisitions of speech technology startups.
Internally, Meta is also pushing employees to adopt AI more widely. Staff are being encouraged to use agentic tools and undergo skills assessments, including system design tests and “vibe coding” exercises. While the company frames these as training initiatives, some employees view them as potential precursors to workforce restructuring.
