Meta CEO Mark Zuckerberg may soon have an AI clone designed to interact with Meta employees and provide feedback, according to a report from the Financial Times cited by The Verge. The reporting says Meta is training an AI avatar on Zuckerberg’s image and voice, including his mannerisms, tone, and public statements—aimed at creating a channel for employees to feel more connected to the founder through interactions with the clone.
What Meta is building
Meta is reportedly building an AI avatar of Zuckerberg that can be used in workplace interactions. According to the Financial Times report cited by The Verge, the company is training the AI clone on Zuckerberg’s image and voice, as well as attributes intended to mirror his communication style: mannerisms, tone, and public statements. The stated goal of this training is “so that employees might feel more connected to the founder through interactions with it.”
This framing suggests the project focuses on generating multimodal outputs—voice and image—that mirror a specific individual’s conversational behavior across multiple interaction dimensions, rather than a generic text-based assistant format.
Zuckerberg’s involvement and related projects
Zuckerberg is involved in training the AI avatar. According to The Verge, he has also started spending five to 10 hours per week coding on Meta’s other AI projects and participating in technical reviews. This places the reported avatar effort within a broader pattern of hands-on involvement in Meta’s AI development.
The Verge also notes that in March, The Wall Street Journal reported Zuckerberg is creating an AI agent of himself to help him complete tasks. This project is reportedly separate from Meta’s work to create an AI avatar of its CEO.
From a technical perspective, these represent different applications: an AI agent is typically oriented toward completing tasks or orchestrating actions, while an avatar is oriented toward interaction—potentially including voice, image, and conversational style. The reporting indicates Meta is exploring multiple roles for AI systems built around the same individual: one focused on task assistance and another focused on interactive representation.
Broader context within Meta’s AI efforts
The Verge references that Meta is also working on a new AI model called Muse Spark, situating the reported avatar work within a wider AI push at the company.
The timing suggests a convergence of two trends in large AI organizations: building systems that can generate or converse, and experimenting with identity-linked interfaces that can make interactions feel more personal. If Meta’s avatar is trained on image, voice, mannerisms, tone, and public statements, it aligns with an industry direction where AI outputs are increasingly tailored to specific user experiences rather than remaining generic.
Potential workplace applications
The reported purpose—enabling interactions so employees might feel more connected to the founder—highlights a potential workplace application for AI systems: using an AI clone as a communication interface inside an organization. This could matter for how companies think about internal communication technology. If an AI avatar can reliably mirror a specific person’s communication style, it may create a mechanism for scaling that person’s presence within an organization.
The reporting does not provide technical details about the model architecture, evaluation methods, or how Meta plans to handle cases where the clone’s responses differ from expected messaging. These implementation details remain unclear from the available information.
Source: The Verge