CALIFORNIA: Mark Zuckerberg has a new side project: a digital version of himself. Meta, the social media giant he founded and runs, is reportedly developing an AI clone of its chief executive, trained on his mannerisms, tone, public statements and views on company strategy, so that staff who cannot get face time with one of Silicon Valley’s most powerful figures can at least get the next best thing.
According to news reports, the rationale is connectivity. With nearly 79,000 employees spread across the globe, not everyone gets a seat at the table with Zuckerberg. An AI facsimile, the thinking goes, could bridge that gap, answering queries and making workers feel a little less remote from the top.
This is not Zuckerberg’s first foray into self-digitalisation. In 2022, he unveiled his own avatar inside his much-hyped metaverse, only for it to be publicly ridiculed for looking like a low-budget video game character. He posted an upgraded version shortly after. The metaverse dream has since been quietly shelved, but the obsession with AI-generated personas has not.
The 41-year-old, whose personal fortune is estimated at more than $220bn, is said to be actively involved in training his animated AI counterpart, which is being built using his image and voice. Meta believes the experiment could serve as a template for influencers and creators grappling with the idea of digital avatars representing them online.
The Zuckerberg clone is just one piece of a broader internal AI push. The Wall Street Journal has reported that a separate “chief executive agent”, a personalised AI system, is already helping Zuckerberg retrieve internal company information faster. The firm is also pressing AI into service across its organisational structure, flattening hierarchies and cutting bureaucracy. “We’re elevating individual contributors and flattening teams,” Zuckerberg said in January.
The commercial logic is hard to miss. Meta is pouring billions into AI in a race against rivals equally hell-bent on dominating the technology. Zuckerberg’s stated ambition is nothing less than “superintelligence”, a system capable of outperforming humans across any cognitive task.
Last week, the company launched Muse Spark, an AI model it claims can estimate calories from a food photograph and plan a family holiday, handling everything from itineraries to child-friendly activity searches, all at once. It has drawn praise for language and visual understanding, though it lags rivals in coding and abstract reasoning.
The AI Zuckerberg, of course, will never face the kind of awkward moments the human one has. In 2023, two days after announcing the layoff of 10,000 staff, Zuckerberg sat before thousands of rattled employees fielding uncomfortable questions about job security and remote working. No AI avatar, however sophisticated, is likely to be put in that particular hot seat.
Then again, that may rather be the point.