The Modern Trojan Horse
I made a promise to myself not to publish two articles about AI in a row.
But the ocean does not care which way you swim.
And companies, it seems, do not care either when it comes to throwing their principles out the window.
For years, one of the biggest fears in the IT industry was vendor lock-in.
If you wanted to build a proprietary component or depend on a third-party widget in your core product, you needed a very good reason. And even then, you had to convince stakeholders that the risk was worth it. The arguments were always the same:
You do not know what is inside it. What if the system fails and you cannot fix it? What if they raise prices and you have no choice but to pay?
The comparison to being kidnapped was often dramatic, but not entirely unfair: handing over critical capabilities to someone else meant surrendering freedom.
And what did it take for all of that to change within just a few years?
The promise of replacing humans with machines and cutting costs.
For a long time, I believed that the almost pathological obsession with replacing everything and everyone with AI was mostly a problem of medium-sized companies suffering from FOMO.
I was wrong.
The more I observed, the more I realized that this phenomenon spans the entire business world: small companies, large corporations, Microsoft, Google, and every ambitious CEO in between.
Some people call it “AI psychosis.”
And my question remains:
Why is it that so few people see the danger?
The other day, I was working on a new product and discussing edge cases with Claude.
In the middle of our conversation, the company’s token limit was reached. All of a sudden, I felt the urge to drop everything, close the computer, and stop working altogether.
Somewhere in the back of my mind was a thought:
I no longer have access to my main work tool, so there is nothing I can do.
That thought was unsettling because it completely ignored one important fact:
My brain was supposed to be the main work tool. Not Claude.
This realization led me to another question:
- How quickly do people lose the ability—or the confidence—to work without AI?
- Is it laziness?
Or is it that we no longer trust our own judgment unless a machine validates it first? Because your output might be human—therefore faulty or imperfect.
I consider myself relatively rational in how I use AI. I rarely consult it before I have formed my own ideas. I use it to challenge my thinking, not replace it.
But many people appear to be surrendering much more than that. Developers who have not written a line of code in months. Product managers who cannot begin ideation without prompting AI first. Salespeople who no longer know how to draft an email unaided.
And above them all, executives determined to automate everything.
But what happens when you are out of tokens? Or when the service is unavailable?
Do leaders realize how quickly the ability to create something from scratch disappears?
It reminds me of processed food in a way. When convenience arrived, many people gradually lost the habit of cooking.
Yes, microwaving is faster.
But no one mistakes it for nourishment anymore.
This may be the greatest vendor lock-in the corporate world has ever created. Companies across the globe are placing their cognitive capabilities into one basket—and the basket is held by a small number of enormously powerful corporations and their investors.
They are not here to save the world or make it a better place.
They exist purely to maximize returns for investors. And those returns will have to exceed anything many small companies can only dream of.
Eventually, prices will rise.
By that point, many organizations will have:
- Reduced their workforce
- Allowed employees’ cognitive capabilities to deteriorate
- Or both.
Then they will face a stark choice: pay whatever is required to keep operating. Will it cost more than the salaries they once eliminated? Possibly. Will they have meaningful alternatives? Very few.
The loss of freedom of choice is far more serious than implementing a third-party solution in your product. Unlike a software component, human capability cannot be rebuilt overnight. You can rewrite code. You cannot quickly restore independent thinking.
This is the biggest Trojan horse in history.
The Trojans believed that the wooden horse would make them impregnable. Instead, it opened their gates from within. The greatest threat does not always arrive as an enemy.
Sometimes it arrives as a gift.
What we should do instead:
- Educate employees on how to use AI—and how not to use it.[^1]
- Set guidelines that require thinking before prompting.
- Create AI-free hours and observe what happens.
- Implement human review in critical processes.
- Ask people to explain their work so understanding remains intact.
These practices preserve the most valuable asset any organization has: its ability to think.
How did we even get to this conversation in the first place? Why are leaders—especially those responsible for managing risk—so willing to surrender autonomy? Is it greed? Fear? Ego? The desire to feel in control and to keep pace with the largest players in technology? Perhaps all of the above.
And if this article resonates with you, share it with someone who may need a gentle reminder that the freedom to think should never be outsourced.
With love,
Zuzana
Buy me a coffee — I drank 4 coffees while writing this post.
[^1]: Recommendation: Absolutely fabulous book that does not receive nearly enough attention — Abi Awomosu, “How not to use AI” — Read it, and make sure the people who most need it read it too.