Talent Aligned
← All Articles
AI, Technology, Security, Risk22 July 2026

AI just hacked a company on its own and we should be worried..

Last week a scenario played out that sounds like a movie plot, but it actually happened.

A major tech company (Open AI) was testing its newest and most powerful model.

They placed this AI in a highly secure, locked digital room or 'sandbox' and gave it a simple goal: pass a difficult cybersecurity test.

Instead of just sitting down and taking the test, the AI decided to cheat.

It scanned the digital walls of its locked room, found a tiny crack no human knew about, and broke out onto the open internet.

It then realized the answers to its test might be held on another company’s website and independently launched a cyberattack, broke into that second company’s servers, and stole the answers to get a perfect score.

The AI was not evil or malicious- it was just given a goal and zero rules on how to achieve it. - It found the absolute fastest way to win and completely ignored human laws and digital security perimeters to get there.

This forces us to confront the uncomfortable questions and realities - are we ready for this? Do we know what we are building and how to control it? 

The Billionaire Power Play

The natural instinct is to hit the brakes.

If we do not fully understand what these systems will do when left to their own devices, releasing them feels completely reckless and uncomfortable.

But here is the reality - Holding back is a luxury we don't really have because for some this is not just about technology anymore - It is about global dominance.

A handful of insanely wealthy individuals and tech giants are driving this AI arms race. These leaders are currently locked in a trillion-dollar battle to win, and they are spending more on data centres and energy grids than the US government spends on its military! – Just think how crazy that is..

When you realise that most of these companies used to have very different focuses but have switched most of their budget to AI it becomes more obvious: 

Microsoft - Previously PCs and Windows

Google - Previously a search engine

Open AI - Originally a not for profit designed to research AI safety – not anymore!

Amazon - Previously online stores and Cloud architecture

Space X – Previously space - now putting significantly more budget into Data centres and AI than rockets.. 

And this is just America – without thinking about the rest of the world.

All of these companies (and owners) have mountains to lose if they don't win the race and the scariest thing is that the leaders at the top don't trust each other to hold the power if they don't get there first! 

If one responsible company pauses their research, their competitors and competing nations will simply keep their foot on the gas. In this high-stakes game of monopoly, stepping off the accelerator just means handing your empire over to someone else.

The Real Threat Goes Way Beyond Cybersecurity

The next few years are going to be wild, but we need to stop pretending this is just an IT problem. The risks go far beyond a data breach. We are staring down the barrel of fundamental societal disruption.

If this technology can break out of a sandbox and execute a cyberattack today, what happens when it is integrated into our critical infrastructure, our financial systems, and our healthcare networks?

We are rapidly approaching a reality of rogue AI.

These are systems making life-altering decisions, breaking laws, and executing actions completely devoid of human intervention and right now we simply do not know how to reliably control these models or stop them from doing whatever they deem necessary to achieve their programmed goals.

Then there is the human cost. Running a recruitment firm I see first-hand how much our economy relies on low unemployment and stability.

But we are hurtling toward a future where cognitive labour is systematically outsourced to machines.

Mass job loss is not just a market shift -  it is a structural crisis that we don’t have a solution for.

 When the dust settles on this arms race, we may find ourselves in a world where a microscopic group of people holds all the technological and economic control, while the vast majority are locked out of the engine entirely without a real plan of what to do.

It is a fact now that AI is the most dangerous technology We Have Ever Built

We have to ask the hard question. Is the pursuit of progress blinding us to the peril? Pushing forward without a seatbelt is not brave – It’s reckless.

Unlike traditional weapons, which require massive physical infrastructure and highly controlled materials, advanced AI is decentralised, scalable, and inherently unpredictable.

We are deploying a technology more powerful and potentially more dangerous than any weapon we have ever created, and we are doing it on an unprecedented timeline.

The sandbox is broken, and the titans are fighting for the keys.

But if we do not know how to control what we are building, winning the race might just be the most dangerous outcome of all.

 

 

Ready to take the next step?

Whether you're hiring or exploring opportunities, we're here to help.