Meta spins up AI version of Mark Zuckerberg to engage with employees

Meta is building an artificial intelligence version of Mark Zuckerberg that can engage with employees in his stead, as part of a broader push to remake the Big Tech company around AI.

The $1.6 trillion group has been working on developing photorealistic, AI-powered 3D characters that users can interact with in real time, according to four people familiar with the matter.

The company recently began prioritizing a Zuckerberg AI character, three of the people said.

The Meta chief is personally involved in training and testing his animated AI, which could offer conversation and feedback to employees, according to one person.

They added that the character was being trained on the billionaire’s mannerisms, tone, and publicly available statements, as well as his own recent thinking on company strategies, so that employees might feel more connected to the founder through interactions with it.

The effort, which is at an early stage, is separate from Zuckerberg’s project to build a “CEO agent” to support him in his role, for example by retrieving information quickly. That idea was first reported by The Wall Street Journal.

Zuckerberg has gone on a multibillion-dollar spending spree over 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. Wall Street investors welcomed the release, with Meta’s shares rising 7 percent on the day.

Zuckerberg has become increasingly hands-on as he oversees Meta’s AI push, according to people familiar with the matter. He has spent five to 10 hours a week coding on different AI projects at the company and sitting in on technical reviews, said one person.

In September 2023, Meta launched its Meta AI assistant as well as a range of AI-powered chatbots exhibiting different personalities based on celebrities such as Snoop Dogg, who agreed to have their voice and likeness used in the feature.

The so-called AI characters were developed as Zuckerberg noted the success of AI companion start-up Character AI, particularly with younger users, according to several people familiar with the matter.

Meta later rolled out an “AI Studio,” which allows users to generate their own AI characters, or creators to build AI versions of themselves to chat with fans.

However, the persona efforts faced controversy last year following reports that users were generating overtly sexual characters, amid concerns from the public and regulators over child safety. Since January, Meta has restricted teen access to its AI characters.

According to people familiar with the matter, Meta’s newly formed Superintelligence Labs have explored a fresh set of characters.

The company has focused in part on making photorealistic embodiments of virtual AI characters, four people said. But scaling the effort has been difficult as the technology requires lots of computing power to achieve realism and avoid a lag in interactions with users.

Meta has also been working on improving voice interactions with the characters. Last year, it acquired two voice companies, PlayAI and WaveForms.

The Zuckerberg character will be trained on images of the chief executive as well as his voice, one person said. If the experiment is a success, influencers and creators might one day be able to do the same, the person added.

Meta has been pushing employees to use AI technology internally to streamline processes and become more efficient. Employees are being encouraged to use agentic tools from the open source software OpenClaw and design their own agents to automate tasks.

Product managers are being invited to do an AI-focused “skills baseline exercise,” according to several people familiar with the matter. This includes a technical system design test, as well as an exercise in “vibe coding.”

Some staff fear this could be a prelude to job cuts. Meta said the exercise was not mandatory and designed to establish where product managers might need extra training and development.

Additional reporting by Cristina Criddle in San Francisco