The Most Powerful Voices in AI Are Warning Us. Why Aren't We Listening?

The defining challenge of the artificial intelligence era is not the technology itself, but that the geopolitical and economic forces accelerating its development are gathering momentum faster than the governance structures needed to shape its future.

For the first time in human history, intelligence itself is becoming an asset that can be owned, scaled and concentrated. The ability to analyse, decide, optimise, persuade and discover is increasingly being transformed into infrastructure. As artificial intelligence becomes more capable, the organisations controlling it are positioning themselves to influence how knowledge is generated, capital is allocated, scientific discoveries are made and strategic decisions are formed.

Many of the strongest signals about the significance of this moment are coming from the very people building the technology.

In 2023, executives and senior researchers from the world's leading AI laboratories issued a public statement declaring that:

"Mitigating the risk of extinction from AI should be a global priority alongside other societalscale risks such as pandemics and nuclear war."

Among the signatories was OpenAI chief executive Sam Altman.

The statement was unprecedented. The leaders of the most powerful AI companies in history publicly compared the implications of their own technology to civilisation-scale challenges. Since then, the signals have only intensified. Altman has called for new governance frameworks. DeepMind founder Demis Hassabis has urged international coordination. Anthropic chief executive Dario Amodei has highlighted rapid economic transformation. Geoffrey Hinton has encouraged governments to consider the implications of systems whose behaviour may eventually exceed our capacity to fully understand. Elon Musk has framed the issue more starkly, describing artificial intelligence as "far more dangerous than nukes".

Collectively, these observations point to a simple conclusion: artificial intelligence is emerging as a strategic capability rather than merely a commercial technology.

The significance of these warnings lies as much in the incentives surrounding AI as in the technology itself. The builders increasingly recognise the scale of the transformation underway, while the geopolitical and economic forces surrounding artificial intelligence continue rewarding acceleration.

That is the defining strategic issue of the AI era.

Throughout history, power has flowed from industrial capacity, financial capital, military strength and access to resources. Artificial intelligence introduces a new dimension because it creates the possibility of concentrating intelligence itself. Advanced systems increasingly generate strategic insight, accelerate discovery and support decision-making across domains ranging from finance and medicine to intelligence, cyber operations and defence.

The economics of frontier AI naturally favour concentration.

The computational infrastructure, energy requirements, specialist talent and capital intensity required to compete at the frontier increasingly favour a remarkably small number of organisations. Concentration is becoming a structural characteristic of the industry rather than a temporary market condition.

The world is unlikely to support hundreds of globally significant frontier AI platforms. It may ultimately support only a handful of organisations possessing the computational infrastructure, capital, talent and data required to compete at the frontier.

Those organisations will become the intelligence layer underpinning economic productivity, scientific discovery, cyber capability, military effectiveness and governmental decision-making.

The logical destination of the current AI race is a world in which a small number of organisations exercise extraordinary influence across economic activity, scientific advancement, information flows, military capability and political decision-making simultaneously. Their platforms may become the intelligence infrastructure upon which governments, militaries and economies increasingly depend.

The executives, boards and investors controlling those platforms may ultimately command a degree of strategic influence that rivals - and in some domains exceeds - that of nation states themselves.

This evolution is unfolding against the backdrop of intensifying great-power competition.

The United States and China have both concluded that leadership in artificial intelligence will influence economic growth, military effectiveness and geopolitical influence for decades to come. Artificial intelligence has therefore become a central arena in the strategic competition between the world's two most powerful nations.

At the same time, a parallel competition is unfolding within global capital markets. Global capital markets are increasingly rewarding scale, capability and dominance. Trillions of dollars in market value now depend upon maintaining leadership in artificial intelligence, creating powerful incentives for firms to accelerate development, consolidate market position and extend technological advantage.

These forces reinforce one another.

Strategic competition attracts investment.

Investment accelerates capability.

Capability attracts users.

Users create dependency.

Dependency strengthens concentration.

The result is a powerful feedback loop increasingly favouring a small number of actors.

Recent events in the White House provide a striking illustration of the challenge. Last week, reports emerged that a proposed executive order introducing additional AI oversight measures was shelved following concerns that new guardrails could slow American progress in the race against China. Whether one agrees with that decision or not, the episode exposed the strategic dilemma at the heart of the AI era. Governments increasingly find themselves balancing governance against competitiveness, oversight against innovation and caution against strategic advantage. The incentives driving the AI race increasingly favour acceleration.

This is how races to the bottom emerge.

Companies pursue market leadership.

Governments pursue strategic advantage.

Investors pursue returns.

Each participant responds rationally to the incentives before them.

Collectively, however, these incentives create outcomes extending beyond the control of any individual participant. The result is a system in which acceleration becomes increasingly attractive, concentration becomes increasingly likely and governance becomes progressively more difficult.

History provides valuable guidance.

The twentieth century demonstrated that transformational technologies benefit from governance frameworks extending beyond commercial interests and national competition. Nuclear, chemical and biological technologies each inspired international agreements because global leaders recognised their capacity to reshape the international order.

Artificial intelligence differs fundamentally from nuclear technology. Nuclear weapons concentrate destructive power. Artificial intelligence concentrates intellectual, economic, informational and strategic power.

Yet both technologies raise a remarkably similar question:

How should humanity govern capabilities capable of reshaping the balance of power?

That question grows more urgent with each passing quarter.

The international community has an opportunity to establish governance mechanisms while the distribution of power remains fluid, and before market structures, geopolitical dependencies and technological concentrations become deeply embedded.

For that reason, global leaders should begin actively developing the foundations of an AI

Governance and Control Treaty - or a comparable international framework - designed to preserve strategic stability, promote transparency, and establish meaningful oversight of frontier AI systems while the future remains open to influence.

Such a framework would support innovation while ensuring that one of the most consequential technologies in human history develops within governance structures commensurate with its significance.

Most importantly, it would recognise a reality that is becoming increasingly clear: artificial intelligence is evolving into strategic infrastructure, with implications extending well beyond the interests of any individual company or nation.

History rewards societies that recognise strategic inflection points early and respond with foresight. The emergence of artificial intelligence as a form of intelligence infrastructure represents one of those moments.

The architects of the AI revolution are already warning that the window for shaping this future is narrowing, their message reflecting an understanding that governance frameworks are most effective before power becomes entrenched.

Great powers continue pursuing advantage.

Capital continues rewarding acceleration.

The race continues to gather momentum.

History may ultimately record this period as the moment humanity decided whether intelligence would become a broadly shared capability or the most concentrated form of power ever created.

The leaders who shape the governance architecture of artificial intelligence will do far more than guide a technology. They will help determine how power is distributed in the twenty-first century.

And once intelligence becomes infrastructure controlled by the few, no amount of hindsight will redistribute it.