Listen to Steve read this post below (7 min audio)
Mark Zuckerberg’s recent announcement of him creating an AI version of himself shouldn’t really surprise anyone. He’s quite obsessed with a few things which have made him one of the richest people in the world. Those obsessions include control over his business (Meta), and the virtualisation of all human connection.
Control: When it comes to control, Zuckerberg has total voting control of Meta (a US$1.71 trillion corporation). Despite only holding 14%, he has 61% of the voting rights and can never be usurped.
Virtualisation: The average internet user spends 6–7 hours of their waking life (around 37% of their life) on screen, and 31% of all their online time on one of Mr Zuckerberg’s platforms.
While I’m not sure this has worked out well for humanity (especially teens), it has worked out well for Uncle Mark, who currently has a net worth of US$239 billion. And despite the $88 billion failure of the Metaverse, the idea to create an AI digital twin of Zuckerberg has legs (See what I did there?)
Previous AI Quiz: Most wrong. AI was on the right.
The Zuckerberg AI Clone
The short summary of his plan goes a little like this: Mark Zuckerberg is developing a photo-realistic AI clone of himself. It will facilitate interactions with and provide feedback to employees. It is to be trained on his voice, mannerisms, and even his strategic thinking. According to reports, his 3D avatar is designed to increase efficiency and employee connection (hmm, really?), by handling meetings and routine tasks, reflecting a shift toward “founder mode” which will allow him to remove himself from the day-to-day and focus on AI strategy.
I think it runs deeper. This is the preamble to the world’s first AI CEO. Because the only person Zuckerberg would ever allow to replace him would be ‘himself’.
This is his plan to be the CEO after his death — with the AI virtual posthumous version of himself. Or possibly if he steps down from the role — albeit unlikely.
It raises some interesting questions. Would it be legally possible for a virtual AI CEO? They’d probably do a better job in many cases. Could this then result in the CEO not having to be paid, saving inordinate millions in salary? Or could that salary go to the owners of the AI?
At present, AI can’t be a legal person — it can’t sign contracts, own anything, be sued, or be held accountable. In 2026, corporate law is still built on human responsibility — fiduciary duty, judgment, consequences. When things go wrong, someone has to wear it. Right now, that’s still a human. But I’m not convinced that won’t change. We need to remember that ‘corporations’ were literally invented to be quasi legal entities to reduce human liability. As far as I can tell, it is only a matter of time before AIs get human rights to do all of the things I’ve mentioned above which are required to run a company.
AI Quiz:
Your Life, Their Asset
But it goes further. Zuckerberg might be training an AI version of himself, but it’s not that much of a stretch to train an AI version of anyone who uses any Meta platform. All those pictures, thoughts, mannerisms, ideas, value systems which Meta has accumulated in the many posts most people make are — wait for it — the property of Meta. This is all clearly laid out in the updated terms and conditions you never read and said yes to.
Yep — he’ll launch a version of your mum, dad or daughter. The kicker is that you’ll have to ‘pay him’ to have continued communication with your loved ones after they die. The AI version of himself he is creating is a test case to make AI posthumous versions of Facebook, Instagram, Threads and WhatsApp user relationships … he intends to monetise — forever.
Zuckerberg is very into forever — from both a control and customer perspective.
If you think it’s unlikely, remember their record and the atrocities (yes, atrocities) committed in the pursuit of profit. Significant anxiety and depression tied to heavy use; Internal research linking Instagram to declining teen mental health and self-harm; a California jury finding them guilty for addictive design targeting young users; under-13 data collection without consent; ignoring known child-grooming activity in DMs; outsourcing moderation to low-paid workers, exposing them to abuse, rape, murder, and suicide - many reporting severe PTSD; and the live-streaming of real-world terror attacks such as that in Christchurch, turning an atrocity into content. Individually disturbing. Together, it’s a system optimised for attention at any human cost.
Legal Battle Ahead
The most important legal developments for AI are still undetermined — yet simple. We must own the rights to our own personas — in all its forms. Our face, thoughts, actions, voices and every form of likeness need to be copyrighted to ourselves, forever.
And if you thought that lawyers will be out of a job due to AI — then you need to revise that idea … some of the most important work they’ll ever do, is only just beginning.
Keep thinking
Steve.
** Get me into do an AI keynote at your next event and as a blog reader - get 2 nights free at my luxury farmhouse on the Bellarine Peninsula.
Hey Steve,
Thanks for these thought provoking articles.
Perhaps Mark watched a show called Upload and instead of seeing a comedy/parody show, he saw a future profit stream.........
https://en.wikipedia.org/wiki/Upload_(TV_series)
Cheers.
Matt (formerly from Cox Automotive)