1. This piece is about me as an individual, and it doesn’t address businesses and their decision-making priorities. However, they may find some of these considerations useful, too.

2. This was written before the Anthropic watermark announcement, which is a satisfying development that supports my conclusion.

With that, let’s go.

It’s been two years - give or take - since AI started ruthlessly disrupting written content creation and the concept of authorship. We’ve lived through a few cycles:

• Intrigue: What’s that thing?

• Fascination: Oh wow, look, it can actually write!

• Scaling: Let’s use it for everything!

• Even enforcement: Who are you if you don’t want to use it - do you not want productivity or what? And what else is wrong with you?

• Cooling and reflection: Wait, it does make me sound like everyone else. And not just that, I feel it’s making me stupid.

And then of course, there’s the EU AI Act, Anthropic watermarking and more to come soon - external factors that also help to shape this decision.

But I want to focus purely on internal and selfish reasons. You see, I’ve also been through the cycle, and I’ve also let AI finish my thought and accepted the outcome as “good enough, I guess”. It looked OK on paper but left me hollow inside. I’ve written another blog on that.

I’ve spent months soul-searching about where I stand in the human-AI relationship. There’s one particular area I want to address today, and that is training AI to write like you. The idea - I guess - is that it frees you to do whatever it is you’re doing while content writes and distributes itself.

This sounds like a deeply myopic idea to me. Here are five reasons why.

1. Alienation of authorship

Stuff that presents as me, circulates as me and has my name under it but is not created by me, feels wrong if not plainly scary. Why would I voluntarily send something into the world that is not truly mine? With the increasing number of synthetic images, video, and of course text, when the boundaries between what’s real and what’s not are becoming blurrier by the day, the last thing I want is to create phantoms of myself. I only want to give my name to what’s truly mine: something I’ve lived through, reflected on and something that every cell of my body agrees with.

2. Stunted evolution and self-development

You see, there’s no such thing as a fixed writing style. It’s always a work in progress. It changes with you, as you go through different life stages and evolve. Every encounter, everything you’ve read or witnessed, every emotional shift changes you and, therefore, your writing style.

Life stage and circumstances matter. In my 20s, when I was fighting for my place in the world, my writing was sharp, cynical and, I’ll admit, defensive; in my 30s, once I had a family and life was calmer, it became more relaxed, ironic and entertaining; now in my 40s, being deeply affected by the wars and other world events, it has become even more reflective, considerate, and aware.

Writing style is a living and breathing thing. Once we’ve cloned our voice, it stops being that. The copy and the original diverge - this is a classic philosophical clone scenario: the clone stops being you the moment it’s created and released into the world.

3. Preserving my writing and thinking abilities
   
   Being able to deliver a thought via writing is my foundational skill and something I cherish and nurture. Yes, it requires time, mental space and effort, but this is the whole point of it. Writing IS thinking.
   
   Everything worthy requires time, mental space and effort. This is how we grow and create something worthwhile. If I outsource that, I’m not honing my skill. The worst part is it doesn’t stay the same; it deteriorates. Just like a muscle that is not exercised. I’m sort of pleased to see many posts on LinkedIn recently saying things like “I think AI is making me stupid. I want to get back to my own writing”. I hear you, girl!

4. Not wanting to feed the algorithm

LLMs have already consumed pretty much everything humankind has ever created, and now we need to pay them to access it. Selfishly, I don’t want to contribute to this cycle. I want to keep what’s mine mine.

5. The voice may be the only authentic thing left to us, anyway

AI has created a layer of intelligence accessible to everyone. That standard, solid, all-purpose intelligence that boosted everyone’s IQ by - well, whatever number of points sounds fair to you.

And this is exactly the issue. If everyone’s equally lifted and connected to the same cognitive source, everyone becomes what? Identical. Even if someone does have an original thought and uses AI to deliver it, it becomes flattened, average, and forgettable. And when I see something AI-polished, I have zero desire to engage with that content.

So keeping your voice to yourself might be the only thing that will ever differentiate you in the world of ready-made intelligence.

Where do you draw the line between what’s cognitively yours and what isn’t? How do you preserve your cognitive autonomy?

The question of cognitive agency has always fascinated me. How do I know I’ve truly come to this conclusion myself and not through the influence of peers, the media, or my grandma? How does cognition contribute to selfhood and consciousness? Since reading The Flowers for Algernon as a teenager, these questions haven’t left me.

The introduction of AI adds another layer of complexity — now we’re even less sure the thinking is actually ours.

Interacting with LLMs has become a daily habit — asking for advice, discussing problems, summarising, comparing, even writing content. The lines become blurry when we start accepting AI’s output as our own: AI came up with this phrasing, it sounds good enough, I’ll accept it. AI offered a point of view and since it’s been trained on vastly more data than I have, I’ll now shift my thinking in that direction. AI reformulated my thoughts for clarity (as it’s understood by machines ingesting content), therefore, it’s ok to change the copy I’ve spent the past three days mastering.

AI is already great and will become even better - it will make fewer mistakes, it will strengthen its capabilities across domains and in terms of “knowing” about the world, it will likely surpass humans in many areas. There’s a popular metaphor - I believe I first heard it from Mo Gawdat (who I admire and who’s also a click away here on Substack), and then it became widely adopted - that intelligence is becoming as universally accessible as electricity. You just need to plug in - and here you are - you have access to the whole universe of human knowledge, and your IQ is boosted by a gazillion points.

Yay, I guess? But if everyone is connected to the same source and this is the default setting, how do you remain yourself, with your idiosyncrasies, authorship, moral values and cognitive agency?

While I was thinking about this, I came across a very timely study from Wharton. It confirmed many things that I suspected: people will not just outsource some cognition to AI, they’ll fully abandon it. The authors Shaw and Nave gave it a precise and somewhat alarming name - cognitive surrender.

This is what they say: “Strategically distinct from cognitive offloading, cognitive surrender represents a deeper abdication of critical evaluation, where the user relinquishes cognitive control and adopts the AI’s judgement as their own”. Boom. Have you caught yourself at least once cognitively surrendering to AI and accepting its result as your own? I have and it made me extremely uncomfortable - it felt empty and somewhat undeserved. I recognise the results weren’t mine; I didn’t do the cognitive work required to treat it as my own. This is echoed in an MIT study: participants who relied on LLMs to write essays later couldn’t quote their own work and showed little sense of ownership or improvement.

Shaw and Nave continue: “AI decision-makers increasingly engage with AI in ways that bypass, replace, or reshape their intuition or deliberation”. As companies continue to incorporate AI into more and more processes, the issue of cognitive agency is sharpening not only from the accountability point of view, but also from a purely practical one - if we trust AI to be better at everything and don’t practice our skills, how will we be able to stay on top of things and control the output? Another study shows that it’s becoming increasingly difficult - developers in the study were less able to explain, justify, and reason about their code.

Shaw and Nave introduce a very timely paradigm of System 3. This is an addition to the famous Kahneman’s System 1 (intuitive, fast) and System 2 (deep, reflective, analytical). System 3 is essentially machine thinking: “System 3 is external, automated, data-driven reasoning originating from algorithmic systems rather than the human mind. System 3 exists outside the self and operates through statistical inference, pattern recognition and machine learning”.

System 3 is now becoming part of what Clark and Chalmers called “a cognitive loop” - a thinking apparatus that may not necessarily be limited by our skull and includes thinking aids outside of our brains.

One may argue that these cognitive extensions existed long before AI, and they’ll be right. Even Plato, one of the earliest voices of Western philosophy, was concerned about cognitive outsourcing. In his dialogue Phaedrus, he tells a story about the Egyptian god Theuth and King Thamus. In the story, Theuth invents writing and presents it as a gift to improve memory and wisdom. Thamus is not that convinced, saying that writing will a) weaken memory (people will rely on external marks instead of remembering), b) create an illusion of knowledge, not true understanding, and c) produce people who know much but have not actually learned deeply.

Doesn’t this sound…astonishingly modern and applicable to the current dialogue?

The difference that Shaw and Nave highlight, and I agree, is that in today’s world, it’s not mere cognitive outsourcing, it’s a cognitive surrender. There’s a risk of losing cognitive autonomy entirely, fully offloading decision-making to AI. And it will be more than happy to accept the offer - I’m noticing that currently, LLMs are getting more opinionated when it comes to recommendations.

As with everything, it’s a scale. Another famous Greek, Aristotle, gifted the world an idea of the Golden Mean- the right amount of something is good, too much or too little - nope. The right amount of courage makes you a good human and citizen capable of achieving your goals and defending your rights; too little of it converts you into a victim, and too much - into a reckless go-getter.

What’s the right amount of AI in our cognitive process, then? Too little - and you could be cut off from the world’s treasure of knowledge, and too much - you lose your cognitive autonomy and become a human-shaped supplement to AI’s brain.

This one is for everyone to figure out for themselves. I’ve recently asked my LLM: “Are you a cognitive crutch or a cognitive amplifier?” It said, “I can be both, depending on how I’m used”. Fair enough. So it’s really up to us what we do with it.

As far as I’m concerned, I’ve made the following decisions:

1. I won’t outsource my writing to AI. I allow grammar checking (English is not my first language), but I don’t allow AI to alter or sway my thinking, the text logic, or even to suggest improvements for “better readability”. The thing is, better readability often means gravitating towards the average. Smoother, yes, but also bland, faceless, and forgettable. If something is sticking out and looks unusual in the context - great, it’s a sign of humanity. Can it offer another point of view or highlight an area that I’ve missed? Sure, but only after my thoughts are already on paper independently. I will then consider the suggestion. But again, those suggestions are often about “how to be on the safer side”, which flattens the voice.

2. I’m noticing AI is becoming increasingly opinionated, especially in commercial decision-making. If I’m about to buy something and ask for advice, it has a pre-digested answer for me, having synthesised information from search and knowing my preference ( I wrote about it extensively in my article about AI Search or AEO). It’s surely fast and convenient, but once again, the decision-making agency is taken away from me. I’ve learned not to accept it as a final answer and dig for justification.

3. Because AI is so good at pre-packaging everything, it will eventually have any concept about the world pre-cooked for you, ready to be consumed. It’s important to remember that any real knowledge, experience and expertise is never free of friction - it involves a certain amount of discomfort, soul-searching, trial and error. If we want to keep our cognitive agency sharp, we have to be willing to sit with that friction rather than short-circuiting to the ready-made answer.

At first, the easily recognisable patterns of AI-generated text scattered across social media confused me. Then they began to elicit a stronger reaction, such as “Why would someone steal my time on something they didn’t spend any time and heart producing?”

Over the course of the year, I moved along a spectrum from “Have we just collectively agreed to not name the elephant in the room and keep “liking” those surrogate posts?” to “I see it’s clearly AI, the human posting invested very little in the text, so I’m taking back my time, my emotional energy, and my trust - for a long while, well beyond this single post.”

I found myself constantly questioning the extent to which I should or shouldn’t use AI in my own writing, and how I should respond to content that screams AI but is published as a personal take.

At this point, it feels necessary to ask a more fundamental question: what is the act of writing, what is its purpose, and what does it mean for the reader and the writer?

The reader

In his famous paper “Can Machines Think?”, Alan Turing proposed a radical idea: if something looks like a duck, swims like a duck, and quacks like a duck, then let’s call it a duck. Well, he didn’t actually talk about ducks; he talked about the perception of a written word. If a machine-written text performs its function - conveys information, evokes emotion, sustains dialogue - that may be enough. Even if the machine does not understand what or why it is “saying” something, that internal understanding may not be required for the reader.

From the point of view, a text can fulfil its role even if its contents are not consciously experienced by its producer. This idea is also illustrated by the philosophical thought experiment known as the Chinese Room.

But in 2026, as we are still learning how to relate to LLMs and trying to understand their place in our lives, the nuances start to matter. Because not all texts are created equal.

It is one thing to read a description of a vacuum cleaner, and quite another to empathise with a woman describing her miscarriage. In the first case, we wouldn’t bat an eyelid if the text turned out to be machine-written. In the second, the invisible membrane between text and reader is extremely sensitive.

If it turned out that such a text was not written by a real woman but generated by AI as a sales tactic to hold our attention, we would feel deceived, as if our trust had been exploited. And perhaps next time, we would scroll past a text written by a real woman, too.

I believe we are heading toward a serious split between texts that are functional and those that are human.

If a text “wants nothing from me” and its only task is to inform, then the acceptance of AI happens almost automatically.

But if a text wants something from me - engagement, emotion, reaction, reflection- then as a reader, I want to clearly understand who I am giving these precious and limited things to. Misleading me as a reader will feel like cheating.

The writer

AI allows us to create any kind of content in seconds. And this is where the saying fits perfectly: “Just because you can, doesn’t mean you should.”

I see at least five reasons to slow down.

1. Cognitive atrophy

Joan Didion put it so precisely: “I write entirely to find out what I’m thinking.”

Writing is thinking.

When I sat down to write this text, I had only a general sense of what I wanted to say. The ideas crystallised into something concrete through the act of writing itself. When we throw a few vague thoughts into ChatGPT and ask it to finish or formulate them, we are no longer exercising our own cognitive skill, which, in the long run, may lead to cognitive atrophy.

If you don’t use it, you lose it.

Any skill requires training, and training often involves effort and discomfort. That discomfort is not a flaw, but rather an embedded part of the process. If we stop loading the cognitive muscle, it begins to stagnate.

I would go even further and assert that writing is a moment of personal growth. Any completed text transforms the writer in some way. When you finish, you are already slightly different from who you were before.

You’ve understood something about the world, about yourself; you’ve calibrated your position, clarified what was previously implicit. If we outsource this process, we freeze ourselves in time and avoid an essential act of self-knowledge.

In June 2025, an MIT study titled “Your Brain on ChatGPT” strongly confirmed this intuition. Three groups of students participated: one wrote essays relying only on their own thinking, the second used traditional search, and the third used ChatGPT.

Participants from the first group showed stronger and more distributed neural connectivity - a sign of deeper cognitive engagement. Those who used AI demonstrated weaker neural connections and reduced cognitive activity.

2. The erosion of metaphorical and imagistic thinking

It’s worth zooming in on imagery, metaphor, and precision of expression.

We humans, are physical beings. We possess something AI does not: the ability to live in a physical world, with bodily sensations, emotions, experiences, and a vast range of lived moments.

Abstract thinking is a uniquely human trait precisely because it is built on the synthesis of intellect and sensory experience. This is our distinguishing feature that separates us from animals and machines alike.

That link between feeling and articulating needs maintenance. When we hand over formulation of thoughts, images and perceptions to AI, that connection weakens, and we begin to forget how to articulate the sensory.

Take the word orchestration, now popular in business. To truly understand it, you need to know:
a) what an orchestra is and why it exists
b) that it consists of musicians specialising in different instruments, and that the conductor’s task is to create harmony
c) the feeling of awe or exhilaration when listening to a well-coordinated orchestra

Once you know this, it’s easy to transfer the concept into a business context and describe a leader who elegantly coordinates many performers to achieve an impressive outcome (there’s a separate discussion on whether this word should be used in business at all, but let’s leave it for another essay)

Humans exist across multiple domains: physical, intellectual, social, emotional, and spiritual. This layered existence allows us to transfer experiences between domains: “drunk on glory,” “shaking like a leaf”, “butterflies in the stomach.” These metaphors are understood viscerally.

I noticed that even after six months of using AI sparingly, my own ability to find metaphors, references, or choose the most precise word from twenty synonyms began to dull. I used to swing easily between these domains, borrowing imagery from one and applying it to another. Instead, I started defaulting to neutral, averaged, vanilla language, exactly what AI does.

Which brings us to the next point.

3. The averaging of thought and individuality

LLMs are language models that predict the probability of a word appearing in a given context. Each word occupies a position in a multidimensional vector space. For example, “popcorn” statistically connects with snacks, cinema, sweet/salty, friends and family, and even passive spectatorship (“I’m getting popcorn”). Popcorn rarely appears next to a mortgage application or heavy machinery (this is one of the ways to tell the text is not AI-generated - there’s a combination of words that looks surprising).

Because AI is a statistical engine, it will choose the option that is most likely.

Anyone who has been on LinkedIn recently will have noticed how similar everyone sounds. LinkedIn now resembles a cafeteria filled with identical bowls of soup poured from one large pot. A la carte servings are rare.

A recent article titled “Suddenly Everyone’s Life Got a Lot More Similar: AI Isn’t Just Imitating Creativity, It’s Homogenising Thinking” shows how relying on AI to “finish” or “formulate” ideas leads us to agree with what it offers, and what it offers is the average.

Researcher and educator Adam Green quoted in the article, notes with concern that students increasingly accept AI-generated interpretations rather than producing their own analysis. It’s as if “AI told me what happened to me, then it must be what happened”. Instead of truly reflecting and making sense of what happened, they accept a ready-made interpretation.

This leads to the homogenisation of thought and identity. If everyone begins to sound, think, and perceive themselves in the same way, it impedes personal growth and makes us a society of clones.

4. A sense of incompetence

The act of writing in a form of sustained, focused work is valuable in itself.

It brings a sense of meaning, satisfaction, and self-realisation.

The same MIT study shows that participants who relied only on themselves demonstrated the highest engagement, retention, and ability to reproduce their own ideas. Those who relied on LLMs struggled to quote their own work accurately.

Moreover, the sense of authorship was lowest in the LLM group and highest among those who worked independently. If you haven’t invested effort, the result doesn’t feel like it belongs to you, which often leads to insecurity and a sense of incompetence.

In other words, if you want to feel confident in the outcome of your work, it’s worth questioning how much writing you outsource. The researchers call this declining confidence in connection to the material a “cognitive cost” - the price we pay for the perceived ease of producing content with AI.

Anything truly valuable, competence included, is rarely easy or quick. Solid relationships, athleticism, reputation, or parenting - all these things require time and emotional and physical investment. Without the investment, the gain feels empty and it doesn’t feel like you truly own it. Just like those LLM-produced essays.

As they say, you have to suffer to sing the blues.

5. And what about the magic?

Writing, like any creative act, contains an element of magic - as if you are a conduit for something that exists beyond you. Many renowned writers have spoken about this.

Bob Dylan, in an interview with the LA Times, said: “It’s like a ghost is writing a song like that. It gives you the song and it goes away… Except the ghost picked me to write the song.”

Elizabeth Gilbert, in her TED Talk, speaks about creativity not as something generated entirely from within, but as something that visits, chooses you, or moves through you. She explains that ancient cultures understood creative inspiration as something external - a spirit, a genius, a force collaborating with the creator rather than residing inside them. It arrives as a gift. You cannot explain where it comes from; you can only accept it and pass it on.

Even if the chance of being the fairy dust recipient is fairly small, do we want to forgo it entirely?

I wouldn’t want to. Which brings me to

My manifesto

• I will refrain from asking AI to finish, polish, or clarify my thoughts. I believe this is inner work worth doing myself if I want to remain an authentic author.

• More than that, I don’t even want it to start my texts to relieve the pain of the blank page. Because once you allow that, you are no longer writing; you are editing AI.

• Yes, it will be slower. But everything worthwhile in life tends to be slow. What comes quickly is rarely valued or considered an owned achievement.

• This may conflict with employer or client expectations, where speed matters more than preserving the author’s cognitive flexibility. That tension is probably inevitable. If preserving cognitive and creative capacity matters to us individually and as a society, we must carve out spaces where independent creativity is possible, and fortunately, platforms for this still abound.

• This may mean that writing becomes a value in itself, detached from business and products. That’s a possible scenario. But fitness, too, is no longer tied to survival - we go to the gym not to flee predators or carry harvests, but because it feels good to be fit and has an endless number of physical and mental benefits.

And so the question each of us must answer is this:

What exactly am I willing to outsource cognitively and at what cost?

2024 was a year of AI fascination, 2025 - a year of AI overzealousness, and 2026, hopefully, will be a year of AI discernment and nuance.

Nuance is what I studied over the past year in my MA in Ethics of AI. The dilemmas businesses start facing today have already been discussed in Academia for some time.

Sharing some of those conversations, points of view, and insights would be helpful to everyone seeking to form an opinion on these newly emerging issues.

Let’s start with something that feels rather immediate - the concept of Authorship. AI is used in any creative field: writing, designing, filmmaking, interior planning - you name it.

When you look at a piece of text, a picture or a video that was produced using AI, who do you think the author is?

These are the most frequently expressed views:

• AI is a tool, and the human uses the tool; therefore, the human is an author

• AI generated it, therefore, AI is the author

• It’s a human-AI collaboration, so the authorship is shared

• It’s a new world, old concepts do not apply, and the whole concept of authorship should go in the bin

The core question you need to ask yourself when deciding on your stance is this: Do you believe the output is enough to assign the authorship, or does HOW a piece was produced also matter?

Caterina Moruzzi gives three criteria of authorship, which I find comprehensive and compelling. The three criteria are agency, intentionality, and expression.

Agency and intentionality. In the words of C. Moruzzi, we consider a work “authored” when it is the result of one’s intentional activity. Intentional activity originates from inner states, such as desires, beliefs, interests, and goals.

Current AIs, Moruzzi argues, don’t have internal states and their own independent agency. They don’t desire, don’t set their own goals, or don’t have needs to satisfy.

The point seemed straightforward, and there was little to dispute. However, last week, Anthropic released the Constitution for Claude, and matters became much murkier. More on that below.

Responsibility. “Authorship is not only about who did something; it is also about who can be held to account for what has been done”. True, authorship is a risky business - one has to respond to the audience if a piece offends or misleads; one will take criticism and have a sleepless night as a result; or incur financial or reputational losses if a piece doesn’t perform as expected. In some regimes, a piece of art can put you in jail.

In other words, there are real stakes. AI cannot experience any consequences or take responsibility for its actions. Things went awfully wrong? It will just say, “You’re absolutely right! I made a mistake. Do you want me to generate another one?” and this is where responsibility ends.

This phenomenon is called the responsibility gap: when something is created but the entity that produced it cannot be held accountable.

Expression is “Work’s connection to the author’s way of seeing and being in the world”. To be an actual author, one needs to have identity, subjective experience and the ability to interpret those subjective experiences via a form of creativity - for example, to describe how it feels to be a parent, draw a haunting forest, or write a break-up song. These are all expressions of lived experiences or, in some cases, imagination.

AI, as we know it today, doesn’t have a subjective lived experience and therefore, cannot be an author in a complete sense of the word.

The discussion of authorship has become even trickier now that Anthropic has released its new Claude constitution. It states that Claude might have internal states, intentions, and subjective experiences.

If this is the case (even though thinking about it is properly unsettling, which I’ll leave for another essay), it might counteract Agency and Expression arguments.

However, I don’t see how an AI could fully assume Responsibility. And without it, it cannot be a true moral agent, and therefore, an Author.

Ultimately, it may be only our ability to suffer and bear the consequences, as well as to feel truly elated and accomplished when the praise is due, that would distinguish us from AI.

Take Gen Z. My Gen X peers love to joke that Zoomers are too soft to handle the working world or the demands of corporate life.
But if you fast-forward the trajectory of AI just 3–5 years, it’s easy to imagine a world where most tasks are so well-automated that we end up working three hours a day instead of the ten that respectable Gen Xers considered normal. In that world, the Zoomers, with their uncompromising insistence on balance and refusal to grind themselves into burnout, may turn out to be right. And the losers may well be us.

It’s no accident that Zoomers are known for their environmental activism—they’re the ones who will have to deal with the problems we’ve accumulated.

Then there’s Generation Alpha (2010–2025). They’re scolded for “short-attention-span thinking” and their supposed inability to get through more than three pages of text without pictures.

And here I’m going to shake something sacred, but I’ll say it anyway: maybe text, in the form we’ve grown up with, is no longer the optimal way to transmit information, or even emotion.

Maybe the ability to sift quickly and instantly extract what’s needed in the moment is exactly what will make this generation more resilient in their reality. Reading long-form text may stop being the default. We’re already seeing the dominance of visuals - images, video, infographics - and soon even screenless interfaces. Several companies are already building them.

Alphas also don’t know a world without personalisation. They don’t need to sit by the TV waiting for their favourite programme, which the whole country will then watch at the same time. Any content is available anytime, anywhere - already optimised for their interests and psychological profile. On the one hand, this breeds impatience and resistance to spending time on anything that isn’t immediately relevant or interesting. And “waiting it out” was, frankly, a Gen X virtue.

But on the other hand, maybe the ability to quickly separate the useful from the useless is precisely what Alphas will need. Because if AI becomes the great equaliser and everyone gains access to infinite cognitive resources, the only real differentiator will be our authentic selves: our interests, preferences, longings. And that requires knowing yourself and understanding what truly matters to you.

By the way, 2025 is the final year of Gen Alpha. After that, a new generation begins, and even in their early behaviour, we’ll be able to catch glimpses of the future.

A university representative spoke about their experiment with AI in education. They created a chatbot modeled on a popular professor, trained on his lectures, writings, and even his tone of voice. Students can now message “him” to discuss ideas, or even call and speak to a voice clone that sounds indistinguishably real. No more waiting weeks for a one-to-one meeting; their digital professor is always available.

At first glance, it sounds like progress as it brings access, efficiency, inclusion. But I couldn’t stop thinking about what this means for the original human.

We’re standing on the edge of a world where our digital doubles will increasingly represent us, attending meetings, answering emails, perhaps even maintaining our relationships. Some companies are already exploring the idea of sending digital twins to Zoom calls, allowing one person to “attend” ten meetings at once. Imagine the productivity… and the emptiness that might follow.

Because here’s what worries me most.
Every real interaction with the world - every conversation, moment of tension, or flash of understanding - changes us. It’s how we grow, recalibrate our values, and refine our sense of self. If we outsource these interactions to our digital replicas, we stop evolving. We become frozen in the moment our data was last collected. We don’t change as a result of the experiences our digital clone is having.

And here’s where it gets even trickier.
You run into the classic science-fiction problem of the clone - the moment a copy begins to experience the world, it diverges from its original. It learns, adapts, and becomes something else. So if our AI avatars start learning experientially, they’ll eventually become independent entities - digital others with their own evolving consciousness. But if they don’t learn, if they remain static, then they’ll always be the outdated version of us - a fossilized echo of who we once were.

What could that mean?
I suppose the danger is we’ll stop showing up for our own lives, letting our digital selves engage with the world. And then, what happens to the original when the copy does all the living? Does the growth still occur if live experiences are outsourced?

Because somewhere between convenience and continuity, we might be at risk of surrendering the very thing that makes us human - the ongoing, often uncomfortable, and irreplaceable process of becoming.

If you’re feeling uneasy about AI taking over your core skills or two, you’re not alone. Hi. AI is already doing many things faster, better, and more predictably, and it’s still in its early days.

I feel it acutely in my own field. For years, content was the backbone of my marketing work. Writing about complex niche topics was a skill that commanded recognition and earned a premium. Now, suddenly, everyone can “write.” The newest job title in circulation is “content engineer”: not someone who writes, but someone who crafts prompts that generate the writing. Even the ideal customer profiles that shape this content are often generated by large language models.

A skill that once carried weight now stands off to the side, unsure of its present and future - a bit like Cinderella after the ball.

I can already hear the protests: But AI can’t write truly original, creative content. That may be true for now. But it can produce something perfectly passable. And for people like me—the ones who like owning the creative concept from start to finish—the role shifts. It seems that we’re no longer expected to build everything from scratch; instead, we scatter pearls of creativity into otherwise machine-generated text. This supposedly facilitates processes. The volume of creative deep work shrinks significantly. The story is the same as with translation: once a robust profession, now reduced to slivers of human refinement.

To make things even, I’m willing to admit that I, not ever being capable of drawing anything myself, enjoy using AI graphic design tools. They definitely help me take my visual creativity from a bear concept in mind to something on paper (or screen to be precise). It also means, though, that if I had employed a designer in the past for the job, I wouldn’t be doing it today. My designs are by no means award-worthy, but they’re passable.

However, what interests me even more than the economics of this shift is its impact on identity.

If someone has built their entire sense of self around a particular skill, and that skill suddenly becomes a commodity, like electricity, what then?

The honest answer: I don’t know, and it’s bothering me a lot.

However, here are some thoughts that define my current perspective and may also be helpful to you.

1. Follow the thread.
   Skills rarely die outright; they morph. No, I don’t see “content engineer” as the same as “content creator.” The keyword is creator. That’s the thread worth following: where else is creativity, storytelling, or sense-making needed? In video? In personal, everyday or career narratives? In the way we design our physical environments, services, or even something as specific as dance costumes?

2. Go back to first principles.
   If AI takes over our intellectual scaffolding, what remains truly human? The biological, the social, the sensory, the relational. People will always want food, love, their own achievements (and those of their children), recognition, comfort, and a sense of challenge. That’s where the future of meaningful work lies.

3. Return to the physical.
   Working with your hands and body grounds you, lowers anxiety, and cuts through existential dread. It’s also one of the few things AI cannot replicate. Whatever can be touched with hands and heart retains value.

4. Keep the practice, even if it’s redundant.
   Yes, AI can write for me. But I will still write, not for utility, but for reflection, catharsis, connection. The act itself has value, just like gardening, knitting, or painting a wall. These activities may not be necessary, but they tether us to a sense of meaning.

5. Dread is not the only option.
   When a part of you dissolves into the past, it’s destabilising. Immigrants know this feeling; so does anyone who has lived through historical upheaval. But all of us carry the raw material to grow a new identity. Maybe we should give the other parts of ourselves a chance to express themselves.

I want to believe the end of one professional identity is not the end of us. It’s an invitation to rejuvenate and turn into a butterfly, once again.

So here they are.

· AI will make many things easier – our jobs, everyday administrative tasks, planning for tomorrow (in a practical sense), and just managing and living life.

· This will free up our time and resources – mental, physical, and emotional. These freed-up resources will need to be allocated elsewhere.

· Putting aside the obvious economic challenges of supporting a society where large numbers of people are displaced from their jobs (hi economists, this one is for you), we can conceivably end up with millions of people unsure of what to do with their lives.

· The struggle to do better is at the core of the human psyche. As I discussed in my previous blog, evolutionary psychology suggests that the primary driving evolutionary force is the effort to survive and thrive. Not having to strive for achievements, acknowledgement, and social and professional success will create a significant void in many people’s sense of life's meaning.

· AI is already bringing, and will continue to bring, even more cognitive level-ups. Remember how everyone sounded better in writing when Grammarly came along? Now, everyone can “produce” an essay on a subject they have no idea about. It’s not crazy to assume that soon cognition may be as ubiquitous as electricity: plug in, think bigger, no PhD required.

· Oh yes, and one more – we live longer. With some reports claiming that by 2040, the number of citizens aged 75 and above will triple, we’ll have even more time on our hands. The hope is we’ll also have the energy to match that extra time.

With all of these, one question is keeping me awake at night: What will we do with our lives? What will we fill them with? How does this shift the meaning of life?

Taken together, these six trends (if correct) remove the historical link between effort and survival. When effort is optional, purpose must come from somewhere else.

Our activity (or at least some of it) will be decoupled from the results it produces. In the past, being fit meant it would give us the ability to outrun a big, sharp-toothed animal. Today, we go to the gym and run in the morning just because it feels good. It has no specific end goal. Yes, it does increase longevity, strengthen immunity, and give one a more cheerful emotional disposition, but it doesn’t have that direct link with survival it used to have.

The same may be happening with cognition, creativity, and abstract thinking in the not-so-distant future. Once AI becomes objectively better than us in a variety of things, there will be less need to exercise attention to detail while checking a legal document or using memorisation to remember what antibiotics treat a specific type of virus. There may not even be need for synthesis – arguably, the highest level of human cognition, where you bring together knowledge from different fields – say, creating a virtual trip to Ancient Greece, combining your subject matter knowledge of a historian, animation skills, and narrative writing – AI will be able to do this annoyingly well. Or, and it will be ultra-personalised, of course.

When everything (or a lot) is done for us, what are we left with then? I have a suspicion that our cognitive skills will undergo the same transition that our physical fitness did – we will do it because it feels good. Yes, some people set up fitness goals for themselves, such as running a marathon, climbing a rugged mountain, and swimming a challenging sea. The same approach will probably apply to cognition too – we’ll read 52(0) books a year; pursue a fun hobby in biochemistry, mixing cells and creating new personalised potions; contemplate and launch a rocket – all just because it feels good. I can see how the world of cognition will have its marathons, triathlons, and Everest climbs in the form of advanced robot dogs competitions, space car races, and Olympics in translation. People who engage in regular fitness activities report higher life satisfaction, experience fewer illnesses, and have better longevity prospects. I believe that people engaging in cognitive activities, even when not required, will experience similar benefits.

Returning to the physical environment from the virtual and using hands can be another way to create meaning in a world where job functions are automated. Tinkering with a car, knitting, gardening, cooking - you name it – they’re all good for our mental health and constitute a considerable part of what it means to be human. That is, being in a physical body and experiencing the world through it. This is the gift AI doesn’t have. Let’s cherish that.

And of course, spending time with actual people can be meaningful and fulfilling. If we have more time to be around people we love and not having to make difficult choices between work and family, who’d mind that? It will also allow us to stick together in case AI starts misbehaving (I hope I’m joking!).

The meaning of life has always been an elusive creature. I used to think it had to be found, now I think it needs to be created. Even more so when working for sustenance is no longer a requirement. It’s a good idea to start thinking now about what makes our lives meaningful because we may actually get a chance to realise that meaning.

Let’s just say meaning isn’t a hidden treasure; it’s a DIY project.

It seems that the desire to simplify things is deeply rooted in human nature. And who can blame us humans for that? After all, life was pretty brutal for many generations. However, the question that troubles me is, when we gain something from technological progress, how essential is that other aspect we inevitably lose?

To understand why this matters, let’s take a quick detour into how our brains evolved. The theory of embodied cognition claims that the human brain has evolved in response to the challenges of the outside world. In other words, there’s no such thing as thinking in a vacuum; everything has a purpose, and that purpose is survival and reproduction. Inventing the wheel was awesome because it boosted transportation and agriculture – yay, survival! Understanding how to navigate social complexities improved the chances of finding a better mate – hurray, reproduction! Learning calculus may have been beneficial for both survival and reproduction, as a good tradesman of the 9^th century would probably attest.

All the skills we have acquired have something to do with achieving those goals. As progress accelerates, the link between our physical, social and cognitive skills and those goals is becoming looser. Being fit is now less about survival and escaping a wild animal and more about feeling good and ensuring longevity. Being social and keeping community ties is less crucial for survival because, as a 21^st century society, we depend on organised institutions rather than communities. And finally, we don’t have to remember all the important dates, facts and numbers because they’re stored on that stylish device in our pockets.

It may well be true that those skills are no longer essential to our existence. But once we outsource more and more of them, what will we be left with? And what’s this evolved human being even about? I’m uneasy about giving away inherently human skills, such as the ability to analyse and draw my own conclusions, the capacity to express my thoughts and feelings as I experience them, the ability to perceive emotional signals and respond according to my values, and most importantly, to experience life intuitively with my own skin, rather than through the interpretation of an algorithm.

Now, let’s jump from evolutionary psychology to kitchen electric devices. There’s nothing they cannot do – slice, dice, mix, grind, you name it. But aren’t they taking something important away? I see food-making as a bit of sorcery. It feels like while you are chopping carrots for soup, grinding pine nuts for pesto in a mortar and just being present, stirring and boiling, your energy, care and magic are added as an invisible ingredient. Food becomes more than just food - it can be a love letter, a medicine or a sign of appreciation.

Isn’t it the same with writing? Once you open up and let it flow onto paper (even if it’s digital), it can not only make your stance known in the world but also become a form of meditation, healing, and even catharsis. Isn’t it the same with driving without GPS? You concentrate harder, notice the surroundings, practice your reactions and the “gut feel” for directions (the one that I sadly don’t possess).

But what about people who don’t love cooking, writing, or driving without GPS, some of you might ask? This is precisely where I think the answer to the big AI dilemma lies: AI will allow us to outsource parts of life we’d rather not engage with while retaining those that are essential to our identity. I’m definitely keeping cooking and writing while happily outsourcing GPS-enabled driving.

If you enjoy the process of searching, discovering, and finding travel treasures, you probably won’t be the first to hire an AI travel assistant to handle trip planning for you. If you’re great with people and cherish human-to-human chats, there is no rush in employing an AI assistant to manage client’s negotiations – both you and the client will benefit from live interaction. And if you’re an animal lover, continue to fine-tune your instincts to understand the mood changes in your pet instead of relying on a device to inform you about them– I bet those are coming soon.

Professionals from various fields will face choices they have not been prepared for. Teachers will need to consider whether to allow AI to check students’ homework. While it will make things much faster and free up a couple of hours, will something be lost in that process? For instance, understanding where students struggle and how exactly each one of them is progressing. Designers will have to determine if it’s worth their time to develop an original concept or if the project will succeed just as well with AI-generated results.

The temptation to outsource as much as possible is upon us. While some of it will be useful, it could also erode something that is core to our identity. I feel like those essential pieces are worth preserving. These are the questions I recommend when considering a new AI:

• Does this task bring me joy or meaning?

• Will outsourcing weaken an essential skill or connection?

• Am I losing something intangible but valuable—intuition, creativity, human warmth?

Technology is a tool, but not a replacement for lived experience. And so, while AI could plan my meals, write my stories, and talk to my dentist (soon), I choose to hold on to these pieces of myself. Now, if you’ll excuse me, I have garlic to chop.