There are days when it already feels like you are slightly behind yourself. A message you meant to reply to. An email you opened and never went back to. A decision sitting somewhere at the back of your mind while you move on to the next thing. Work has a way of piling up like this, quietly, without making a fuss. And then there is that familiar thought: if only you could be in two places at once, things would feel easier.
It is not an extraordinary idea. Just a quiet, practical one. The kind that comes up in the middle of an ordinary day, when you realise there is more to do than you can keep up with. It is also the kind of problem technology has always tried to solve.
As Mark Zuckerberg suggested while outlining his vision for AI, the goal is not to automate all valuable work, but to support individuals with intelligence shaped around their lives.
Mark Zuckerberg is creating an AI clone of himself. Recent reports suggest that he has been building an internal AI system to assist him in his role as chief executive. It helps him find information quickly, bring together answers, and move through decisions with fewer of the delays that typically come with large organisations.
The idea of being in more than one place at once
At one level, this sounds like efficiency. The kind that makes work smoother and faster. But it also feels familiar in a quieter way.
Most days are not made up of one big task, but many small ones. Messages, replies, follow-ups, decisions. Things that are left hanging. An AI system, at least in theory, helps close some of those gaps. It responds when you cannot. It keeps things moving.
At Meta, this seems to be part of a larger shift towards AI that feels more personal. Not just tools you use now and then, but systems that begin to adjust to the way you work.
So what would it take to create a version of you?
The question sounds bigger than it really is.
It begins with something you are already doing. Every message you send, every email you write, every decision you make leaves behind a pattern. You may not notice it, but it builds over time.
You tend to phrase things in a certain way. You reply with a certain rhythm. You make similar choices in similar situations. None of this feels planned, but it repeats often enough to be recognised.
As Manish Goyal, a UX designer at Onething Design Studio, says, “What looks like spontaneity is often repetition. The way we respond online becomes predictable over time.”
This is what such systems learn from. Not your personality in the abstract, but your habits in practice. The words you choose, the tone you fall into, the way you respond to certain kinds of messages.
Over time, this can be used to train a system to do something simple but powerful. To suggest what you might say next. Then, to complete that response. And eventually, to send it.
It does not happen all at once. It begins with small assistance. Drafting a reply. Finishing a sentence. Offering options. But as the system sees more of your behaviour, it becomes better at predicting it.
AI does not need to fully understand you. It only needs enough of these patterns to begin guessing what you might do next. And once it can do that reliably, it starts to feel less like a tool, and more like a version of you.
When AI begins to speak in your voice
At first, it feels simple. Helpful, even. A suggested reply here, a drafted message there, a small saving of time. But there comes a point where something shifts. The system is no longer just helping you respond. It begins to respond in a way that feels like you would have written it yourself.
Zuckerberg’s own use of an AI system is still quite practical. It helps him move through information more quickly and make decisions with less delay. But the idea behind it goes further. If a system can recognise how you usually respond, it can begin to respond on your behalf. And in many cases, the difference may not be obvious.
But there is a limit to how far this can go.
What cannot be learned
This is where the idea becomes harder to ignore. These systems rely on your past, on everything you have already said and done. They do not know how you might change your mind tomorrow. They cannot account for instinct, contradiction, or the small inconsistencies that make people human.
As Prachi Saxena, clinical psychologist and co-founder of The Emotional Wellness Initiatives, explains, “A person is not just a pattern of responses. We are shaped by context, emotion, memory, even unpredictability. That is the part no system can fully replicate.”
And yet, in everyday interaction, that difference may not always be obvious. If something sounds like you, follows your way of thinking, and carries your tone, does it matter whether you actually wrote it, or does it simply matter that it feels like you did?
Zuckerberg’s experiment is still at an early stage. There is no fully independent version of a person yet. What exists today is more limited, systems that assist, retrieve, and summarise. But the direction, at least for now, seems clear. Creating a version of yourself does not begin with technology. It begins with behaviour, with the patterns you leave behind, often without realising it. Technology is simply becoming better at reading them.
And once it can, the question is no longer whether a version of you can exist. It is what part of you it is really capturing, and what part of you is still your own.