The news is era-defining: the White House is pushing the world’s largest investment banks [1] — we’re talking about giants like Goldman Sachs, Bank of America, and Citigroup — to use Anthropic’s AI (Claude Mythos) for their cybersecurity. An alliance between government, finance, and big tech that feels like a trailer for a sci-fi movie.
But behind the spotlight of this seemingly futuristic move, two fundamental questions hide that concern us all, from Trento to San Francisco:
- Is AI about to make cybersecurity experts obsolete?
- While the US ties itself to a single “black box,” what is Europe doing to avoid being left behind?
Spoiler: the answers are much more complex and interesting than a simple “yes” or “no.”
The myth of “job-stealing” AI: the data says otherwise
The first instinctive reaction is fear. If an AI can perform red teaming and find vulnerabilities, what will happen to the tens of thousands of professionals who do this for a living today? [1]
Fortunately, the data comes to our rescue and paints a completely different picture. Let’s start with a striking number: according to the 2025 Fortinet Global Cybersecurity Skills Gap Report [1], 4.7 million cybersecurity professionals are missing globally. That’s not a typo. In Italy, the gap is equally serious, with about 136,000 job postings per year in the ICT sector against only 73,000 new talents entering the field [1].
AI isn’t entering a saturated market to steal jobs; it’s arriving in a sector in full emergency where human forces are no longer enough.
The real game: US Monopoly vs. European Sovereignty
According to Gartner [1], by 2026, 50% of Security Operations Centers (SOCs) will use AI as decision support. The goal is not to replace the analyst, but to give them a “superpower”: automating repetitive tasks, analyzing amounts of data unmanageable for a human, and flagging only the anomalies that require intuition and creativity. AI becomes an ally, a force multiplier that allows understaffed and stressed teams to focus on strategy instead of routine.
In my opinion, the security professional’s profile will evolve from a hyper-specialized technician to a strategist and supervisor of intelligent systems. Their task will no longer be finding the needle in the haystack, but teaching the AI how to search for the needle and, above all, understanding if what it found is actually a needle.
USA vs. Europe, Monopoly vs. Sovereignty
And here we get to the geopolitical heart of the matter.
Act One: The American bet on “everything, now”
On one side, we have Team USA. The government acts as a sponsor, and Wall Street adopts a proprietary and powerful solution en masse. Fast, efficient, centralized. The quintessence of the American approach. Even Jamie Dimon of JPMorgan, a known skeptic, was invited to the table [1] (though he did not participate). The risk? A frightening vendor lock-in. Entrusting the security of the planet’s most critical financial infrastructure to a single private provider creates a single point of failure with unimaginable consequences. And, let’s not forget, a historical precedent that teaches us to be cautious when government agencies (like the NSA) and tech giants dialogue too closely.
Act Two: The European marathon for autonomy
On the other side is Team Europe, playing a completely different game. Slower, more fragmented, but with a very clear objective: digital sovereignty. Europe has understood that technological dependence is the new energy dependence. And it’s moving. [1]
Contrary to what people think, these are not just words. The network of European Digital Innovation Hubs (EDIH), with over 200 active hubs, is helping SMEs digitize by choosing European technological solutions. It’s not just about innovating, but doing so while building an autonomous ecosystem. Furthermore, national initiatives like the “Fondo per la Repubblica Digitale” in Italy (with about 220 million euros) demonstrate an investment in closing the skills gap [1].
Two philosophies, one choice: what digital future do we want?
What we are witnessing is not just a technical debate, but a clash between two philosophies. The American approach privileges speed and power, even at the cost of centralizing power in the hands of a few giants. The European approach is more cautious, aiming to build a resilient, open, and distributed alternative, even if it requires more time and coordination.
For those working in the sector, the lesson is twofold. First: AI is not a threat, but the tool that will define the next decade. Ignoring it is not an option. Second: the choice of tools we use is never neutral. Behind software lies a business model, an ideology, and a geopolitical strategy.
The real question, then, isn’t whether we will use AI for our security, but which AI we will choose: the closed and centralized one or the open and federated one? [1]
The answer to this question will decide who controls the keys to our digital future.