Tech Armageddon is here, By Osmund Agbo

Last August, I visited home. Like millions of travelers around the world, I had to book a flight, and I turned to Kayak, the online travel search engine that helps users compare prices for flights, hotels, rental cars, and vacation packages. I entered my travel details, and within seconds, the website spat out multiple itineraries, complete with airlines, departure times, and prices. I compared the options, weighed the costs, selected a flight, entered my credit card information, and completed the purchase.

It was a relatively straightforward process. The problem, however, was that I still had to do the searching, comparing, deciding, and booking, all of which tech bros believe takes too much time for such a simple task. Now, imagine a technologically advanced future in which I would not have to go through all those steps. Instead, I would simply give an instruction to an artificial intelligence assistant like this: “I need to be in Lagos next Tuesday morning. Find a reasonable flight, check my calendar, choose the best option, and book it if it costs less than $800.” Insane, right?

Well, there is no need to imagine such a future. It is already here with us. Meta, Mark Zuckerberg’s company, launched an AI agent called Muse just last month, and the idea behind Muse as a personal AI agent is precisely to make this kind of delegated action possible. Meta’s Muse can potentially search the web, examine your calendar, compare flights, evaluate travel times, recommend an option, seek your approval, complete the booking, and report what it has done. In fact, more ambitious AI agents could even manage correspondence, coordinate schedules, fill out forms, and pursue complicated goals with minimal human supervision.

The significance of these disruptive technologies is not simply that we will spend less time booking flights or writing emails. It is that we are moving from artificial intelligence that responds to our instructions toward artificial intelligence that acts on our behalf. We are beginning to delegate not just tasks, but also decisions and, increasingly, the authority to carry them out.

Meta is just one of several competitors in a global race. OpenAI, the company behind ChatGPT, is developing increasingly capable general-purpose assistants. Anthropic’s Claude has become particularly useful for sophisticated writing, coding, and analyzing complex documents. Google is integrating AI with its vast information ecosystem. Chinese company DeepSeek is pursuing powerful models that can be developed and deployed at comparatively low cost, while Elon Musk’s xAI is building systems connected to real-time information. Each company has its own ambitions, advantages, and commercial strategy.

But this competition extends far beyond Silicon Valley. It has become a contest among nations, with enormous implications for economic power, military capability, cybersecurity, and geopolitical influence. The race increasingly resembles an arms race, with one crucial difference: unlike conventional weapons, artificial intelligence is not confined to military installations or battlefields. It is becoming embedded in the infrastructure of everyday life.

Major breakthroughs in artificial intelligence are beginning to look less like product launches and more like weapons tests. The technologies being unveiled by competing companies promise to transform human existence, but they also raise a frightening question: What happens when machines become so powerful that even their creators begin to fear what they have built? Governments are scrambling to understand these technologies, assess their implications for national security, and determine what protections might be necessary before the genie escapes the bottle.

America is still leading. I looked at the world’s ten most valuable companies by market capitalization and found that eight are American, with Apple and Nvidia leading the pack, each valued at $5.21 trillion. The only two non-American companies on the list are Saudi Aramco and Taiwan Semiconductor Manufacturing Company (TSMC), the Taiwanese semiconductor manufacturer.

These companies are investing more annually in AI research, computing infrastructure, and semiconductor production than some countries spend on their entire yearly budgets. It is obvious that whoever controls the most advanced computing infrastructure possesses a potentially decisive advantage in developing increasingly capable systems.

The United States also has a substantial lead in data center infrastructure, although estimates vary depending on how facilities are counted. The United States has approximately 5,427 data centres, accounting for roughly 45 per cent to 46 per cent of all active data centres worldwide, compared with China’s 449. China, meanwhile, possesses an enormous domestic market, a large population, and significant technological ambitions. Both countries understand that AI could shape the balance of power for decades to come. Neither can afford to ignore the possibility that falling behind could carry economic and strategic consequences.

There is, of course, much to celebrate. AI can accelerate scientific discovery, improve medical diagnosis, help doctors interpret complex information, and make knowledge accessible to people who might otherwise struggle to obtain it. A businessman can use an AI assistant to review contracts, compare proposals, and examine hundreds of pages of regulatory documents in a fraction of the time such work once required. Researchers can explore questions that might previously have consumed months or years. Tasks that once demanded entire teams can increasingly be completed by a single person equipped with the right tools.

For someone like me, who writes, practices medicine, and manages business interests, the possibilities are extraordinary. AI can become a research assistant, an editor, an analyst, and an administrative aide. Properly used, it can free human beings to concentrate on judgment, creativity, and the questions that matter most.

Yet every technological revolution creates vulnerabilities alongside its opportunities. The more deeply we integrate AI into our lives, the more dependent we become on the digital systems that sustain them. Our banks, hospitals, power grids, retirement accounts, government agencies, and communication networks are increasingly connected. This interconnectedness makes modern life efficient, but it also creates opportunities for disruption on a scale that earlier generations could scarcely have imagined.

The growing alarm surrounding powerful AI models illustrates the danger. This month, Anthropic announced that it had built an artificial intelligence model so powerful that it was too dangerous to release widely per New York Times report. The company named 11 US organizations as partners to help mount a defense.

The paper reported that within two weeks, the model, called Mythos, had triggered a global scramble unlike anything yet seen in the AI era. Anthropic has said that Mythos is uncannily capable of finding and exploiting hidden flaws in the software that runs the world’s banks, power grids, and governments. The model had become a geopolitical bargaining chip. Anthropic’s response has highlighted the difficult question of how to distribute powerful capabilities responsibly without placing them beyond the reach of those who could use them defensively.

Imagine waking up one morning to discover that you cannot access your bank accounts. Your financial institution is struggling to recover from a sophisticated cyberattack. The company managing your retirement investments has lost access to critical systems. Hospitals are experiencing disruptions, businesses cannot process payments, and electricity providers are scrambling to protect their networks. Now imagine that these attacks are not isolated incidents but coordinated assaults, accelerated by AI systems capable of identifying weaknesses and executing operations at machine speed.

This is not a prediction that such a catastrophe is inevitable. It is a recognition that the consequences of failure could be immense. AI does not need to become conscious, develop hatred for humanity, or decide to overthrow its creators to cause extraordinary harm. Human beings equipped with powerful AI tools may be sufficient. So might a poorly designed system operating with excessive privileges, inadequate safeguards, or access to critical infrastructure. The danger lies not only in what machines might independently do, but also in what people can accomplish through them.

There are other costs. The expansion of data centers consumes enormous amounts of electricity and requires substantial infrastructure, water, and investment. Communities are increasingly concerned about the environmental consequences and rising energy demand. Meanwhile, the concentration of technological power in a small number of companies raises questions about accountability. Who is responsible when an autonomous system makes a disastrous decision? Who bears the loss when an AI-driven attack destroys a business or compromises a public utility? And who decides which risks society must accept in exchange for technological progress?

These questions have become personal for me in an unexpected way. If the systems underpinning financial markets can be compromised, should I reconsider how much of my investments I keep in stocks, bonds, and certificates of deposit? Should I buy more gold and physical assets instead?

The temptation to retreat into tangible wealth is understandable, but it is not without risk. Gold can preserve value under certain conditions, but its price fluctuates. Physical assets can be damaged, stolen, or difficult to liquidate.

I believe the sensible response is not necessarily to abandon the financial system, but to understand the risks, avoid excessive concentration, maintain appropriate emergency reserves, and ensure that one’s financial life is not dependent on a single institution or digital platform. Portfolio diversification therefore remains important.

Ultimately, the challenge is not to stop AI. That is neither realistic nor necessarily desirable. The challenge is to ensure that the speed of innovation does not outrun our capacity to manage its consequences. Governments must develop credible safeguards. Technology companies must accept that their responsibility extends beyond releasing the next impressive product. Critical infrastructure must be protected against increasingly sophisticated attacks, and the public must be educated about both the benefits and the dangers.

Humanity has always invented tools more powerful than those that preceded them. The difference today is the speed at which these tools are improving, the scale at which they can operate, and the extent to which we are inviting them into the systems on which civilization depends.

The technological future may be magnificent. AI could help us cure diseases, unlock scientific mysteries, and solve problems that have resisted generations of human ingenuity. But if we build systems capable of acting across the world without building equally capable systems of oversight and protection, we may discover that the greatest danger is not that artificial intelligence becomes too human, but that human beings become too careless with the power they have created.

Tech Armageddon may not arrive with explosions, mushroom clouds, or armies marching across borders. It may begin with a silent breach in a server, a compromised network, or a single instruction executed at breathtaking speed.

Osmund Agbo is a medical doctor, author, and the 2026 recipient of the Thabo Mbeki Prize for Scholarship and Public Service.