Is Black Mirror becoming reality? Zuckerberg has been exposed for building AI digital avatars, so in the future—when interacting with employees—you won’t need real people to attend.

Meta CEO Mark Zuckerberg is developing an ultra-realistic AI digital avatar, aiming to enable interaction across time and space. Meta has put a budget of more than $100 billion into this effort to drive its Personal Superintelligence plan.

An ultra-realistic AI doppelgänger appears at Meta, and Zuckerberg invests in training to build a digital double

Meta CEO Mark Zuckerberg (Mark Zuckerberg) is leading a company experiment to “digitalize” himself through artificial intelligence technology. According to a report by The Financial Times, Meta is internally developing an AI-driven version of Zuckerberg with lifelike feel and 3D visual effects. The design goal is to allow the CEO to interact with employees worldwide in real time without having to attend in person. The project is led by Meta’s newly established Superintelligence Labs, and it is currently in an initial testing phase.

This digital double features highly realistic looks, and deeply integrates Zuckerberg’s language style, facial expressions, vocal tone, and public speaking patterns. To ensure the AI’s responses align with the person’s strategic thinking, the development team trained it using Zuckerberg’s recent internal views on the company’s development and his public statements. Zuckerberg has shown a strong interest in the project, even personally participating in technical fine-tuning, so that employees feel a connection similar to speaking face-to-face with the founder when communicating with the AI double.

Compared with the virtual avatar that was criticized on a metaverse platform in 2022, Meta is now pursuing extreme realism. This 3D character can provide feedback, join conversations, and simulate natural video calls in virtual spaces. To improve the naturalness of the AI’s speech and eliminate interaction lag, Meta has already acquired two voice technology companies to strengthen its technical foundation.

• Related news: Even worse than a game from 26 years ago! Meta’s metaverse visuals get slammed as “too outdated”

Shifting to superintelligence, Meta pours huge resources into its Personal Superintelligence plan

As Meta shifts its focus toward generative AI, the company’s capital expenditure budget has surged accordingly. Market data shows that Meta’s capital expenditures in 2026 are expected to be $115 billion to $135 billion, nearly twice the size of the prior year. This spending will mainly be used to expand computing capacity to support the vision of “Personal Superintelligence.” Meta’s recently released Muse Spark model is the product of this strategy—this dedicated model with capabilities for health reasoning and visual understanding has driven the company’s stock price up on the day of its release.

• Related news: Meta’s new AI model Muse Spark is unveiled! Expected to be introduced to FB and IG within a few weeks—Meta AI App jumps into the top five

Meta’s internal AI transformation is reflected in the CEO’s doppelgänger, and it is also being rolled out across all employees. Currently, Meta asks employees to introduce agent tools into their daily work processes, and encourages the use of open-source software to design automated agent programs to improve efficiency. Product managers also took part in a technical evaluation called “Skill Benchmark Practice,” which includes system design testing and the recently popular “Vibe Coding.” Some employees are worried that this could be paving the way for future layoffs, indicating internal unease brought about by technical changes.

In addition, Meta is extending this kind of AI virtual technology to the creator ecosystem. Through the AI Studio platform, in the future, influencers and creators can build their own AI digital doubles to respond to fan messages and handle business interactions. In social commerce, Meta has already promoted new features in multiple countries, allowing merchants to link product catalog connections to Reels, so that content output can directly turn into digital storefronts with shopping functionality.

Digital legacy management patent: Meta envisions users continuing to post after death

Meta’s application of digital doubles has extended to the boundary of life-related rights and interests. A recent patent document shows that Meta is exploring the possibility of using AI to mimic the behavior patterns of deceased users. By analyzing a user’s posts, comments, like records, and private message content during their lifetime, this technology trains a bot that can mimic the person’s personality and tone, keeping the account active after the user passes away.

This “AI copycat” can automatically post updates, reply to private messages from friends and family, and even communicate with the living through voice or video call technology. In the patent document, Meta explains that when an influential user or a close person disappears, it can impact the community’s overall user experience. Therefore, an AI copycat can fill this emotional gap. This is similar to the patent concept for AI chatbots that Microsoft previously obtained, and both involve simulating interactions with deceased individuals.

The technology has triggered intense criticism on social platforms. Netizens have likened it to a storyline from the sci-fi anthology series Black Mirror, worrying that bodies will be commercialized and even become tools for scam groups. Scholars point out that the key to human experience of grief lies in facing a real loss. The superficial “resurrection” of an AI copycat may lead to psychological confusion and even hinder the normal process of mourning. Meta clarified in response that the patent application is to protect the technology concept and does not mean the feature will definitely be formally implemented.

Related news: Black Mirror becoming real? Meta’s new patent “AI copycat” lets you keep posting after you die—sparking major ethical controversy

Ethics and privacy challenges: the double-edged sword effect of digital avatar technology

As virtual doppelgänger technology moves toward implementation, questions about rights and legal issues behind it are gradually coming to the surface. The Zuckerberg AI avatar developed internally at Meta faces ambiguity around the boundaries of power. When employees talk with an AI CEO avatar, they may treat the AI’s responses as official instructions. If the AI generates incorrect suggestions, the relevant accountability mechanisms and legal responsibility are still in a gray area.

Privacy and data security are another major test. Training lifelike doubles requires massive amounts of personal biometric data, including face capture and voice samples. For ordinary users, this involves the right to manage a digital legacy after death; for enterprises, it faces the risk of confidential strategic leakage. Meta has previously drawn attention from regulators due to inappropriate statements generated by AI chatbots, showing that AI agents, in the absence of a well-developed governance framework, can easily trigger social conflicts.

Despite ongoing controversy, Meta remains firmly committed to developing AI agentification. From on-the-ground testing of stablecoin payments at Japanese financial institutions, to Meta researching and developing 3D virtual avatars, the combination of blockchain and AI technology is trying to redefine the boundaries between the real world and the digital world. In the future, when receiving information on social media, the sender could be a real person—or it could be a perfectly trained AI double. This technology shortens interpersonal distance while also challenging the value of authenticity.