AI: A Threat or a Mirror of Our Own Mistakes?

Cybersecurity

Tech

Privacy

Philosophy

AI: A Threat or a Mirror of Our Own Mistakes?

Ecaterina Sevciuc

4 min read·

Just now

--

Introduction

It is hard to open a news feed today without encountering apocalyptic predictions: AI will take our jobs, AI is “hallucinating,” and AI will finally destroy what’s left of our privacy. This has evolved into a form of mass hysteria. However, as a developer with a background in law and banking, I see the situation differently. To me, AI is not the cause of the crisis, but a logical consequence of how we have been constructing the digital world over the last few decades.

The first law of economics is simple: supply follows demand. We live in an era of unprecedented information overload. The sheer volume of data is so immense that the human brain, even a brilliant one, is physically incapable of processing it in time to draw correct conclusions.

In today’s world, time is money, and modern realities dictate that you should have known everything “as of yesterday.” AI emerged at just the right time as an answer to this demand. It is not a threat, but a “super-filter” — a tool that helps us navigate the ocean of information we created ourselves.

For decades, we have lived within an inflated bubble of perceived security. The digital world was built on data: corporations needed user information for profit, leading to the creation of apps, games, and services where registration and the surrender of personal data became mandatory.

All of this was wrapped in the beautiful slogan of “cybersecurity.” In reality, however, this was often mere ”Security Theater” — protecting corporate interests rather than the client. We inflated this bubble ourselves, trading privacy for convenience because “that’s just how it works” now. It is time to break this cycle: we must stop thinking only about monetization and finally focus on real security architecture. And this is exactly where AI can become our strongest ally, helping us design systems that don’t collapse like a house of cards at the first sign of trouble.

In economics, there is a concept describing the sudden collapse of an unstable system — the Minsky Moment. It is the instant when a structure that seemed stable suddenly proves fragile under the weight of its own debts or errors. The more the bubble of illusions was inflated, the more tangible its correction becomes.

The emergence of AI has exposed a massive ”Information Asymmetry”: it turns out that those who promised to protect us often lacked a solid foundation themselves. AI doesn’t “steal” data in the classical sense; it simply extracts and systematizes what is already in the public domain. This accessibility was created by people, not algorithms. Blaming AI for a lack of security is like blaming a librarian for a library having too many bad books. The librarian didn’t write them; they just help you find them.

In “The Matrix,” Morpheus offered a choice between the blue pill and the red pill. Perhaps AI is that very “red pill.” It is uncomfortable to admit that our security concepts have collapsed, so it’s easier to shift the blame onto a “soulless machine” that cannot talk back.

But acknowledging a mistake is the first step toward fixing it. Instead of fighting AI, we should use it as our most experienced consultant. We should ask: “How can we rebuild the system so that protection is real, not just a marketing slogan?”

Perhaps this “librarian” is exactly who we need to finally bring order to our own digital home.

This article was originally written in April, reflecting my core perspective at the time — and today, nothing about that vision has changed.

If anything, time has only confirmed it. Humanity always gets what it builds, and it is far too easy to externalize blame onto “the algorithm.” But unless we start taking responsibility for how we design these interactions, the grim sci-fi scenarios we love to watch on screen will simply become our reality.

We need a fundamental paradigm shift in cybersecurity. An LLM is not just a traditional database or a server — it is a probabilistic, language-driven system interacting directly with fluid human psychology. The old, static guardrails no longer work because we have entered an entirely new evolutionary phase of software engineering.

As the old saying goes: Adapt or go extinct.

This exact shift — moving away from security theater toward deterministic, auditable threat modeling for human-AI interactions — is what inspired me to architect AURA (AI User Risk Assessment Framework). Feel free to explore my open-source security work or connect with me on GitHub!