Meta is cooking up an AI Mark Zuckerberg so the real one does not have to interact with humans.
The $1.6tn group is developing photorealistic, AI-powered 3D characters that users can interact with in real time, according to four people familiar with the matter.
The company has recently prioritised a Zuckerberg AI character, three of those people said. Meta chief executive Mark Zuckerberg is personally involved in training and testing his animated AI, which could offer employees conversations and feedback.
The character is being trained on the billionaire’s mannerisms, tone, and publicly available statements, as well as his recent thinking on company strategy. The idea is that staff might feel more connected to the founder through interactions with his polished digital stand-in.
This effort is early-stage and separate from Zuckerberg’s “CEO agent” project to support him in the role, such as quickly retrieving information. That idea was first reported by the Wall Street Journal, which will be thrilled it is not being replaced by a chatbot yet.
Zuckerberg has gone on a multibillion-dollar spending spree in the past year, promising to develop “personal superintelligence” and catch up with rivals such as OpenAI and Google in building cutting-edge models.
On Wednesday, Meta released Muse Spark, a small, closed “purpose-built” model for use across its products, with advanced capabilities in areas such as health reasoning and visual understanding.
The cocaine nose jobs of Wall Street cheered, sending Meta shares up seven per cent on the day. People familiar with the matter said Zuckerberg has become increasingly hands-on, spending five to 10 hours a week coding on different AI projects and sitting in on technical reviews.
Meta has been playing with personality-driven bots for a while. In September 2023, it launched its Meta AI assistant and chatbots with celebrity-style personas, including Snoop Dogg, who agreed to have their voice and likeness used.
The so-called AI characters were built as Zuckerberg clocked the popularity of AI companion start-up Character AI, particularly with younger users, according to several people. Meta later rolled out an “AI Studio” that lets users generate AI characters or build AI versions of themselves to chat with fans.
That whole persona push ran into trouble after reports that users were generating overtly sexual characters, sparking child safety concerns from the public and regulators. Since January, Meta has restricted teen access to its AI characters.
People familiar with the matter said Meta’s newly formed Superintelligence Labs has explored a fresh batch of characters. The company is focusing in part on photorealistic embodiments, but scaling is difficult because realism requires a lot of computing power, and you cannot have lag when someone is meant to be talking to your boss.
Meta has been working on improving voice interactions, too. Last year, it acquired two voice companies, PlayAI and WaveForms.
One person said the Zuckerberg character will be trained on images of the chief executive and his voice. If it works, influencers and creators might one day do the same, which should go down well with anyone who has ever wanted to outsource their own personality.
Internally, Meta is pushing staff to use AI to streamline processes and “become more efficient”, which is never a sentence that makes workers relax. Employees are encouraged to use agentic tools from the open-source software OpenClaw to build their own agents and automate tasks.
Product managers are being invited to do an AI-focused “skills baseline exercise”, including a technical system design test and an exercise in “vibe coding.” Some staff fear it is a warm-up for job cuts, while Meta says it is not mandatory and is meant to identify training needs.
