The interdiction of Mythos marks the containment of technological chaos

The case

A recent cover story in The Economist brought to the centre of public debate a fact that may appear technical, yet carries profound institutional implications: the suspension of the release of an artificial intelligence (AI) model with an advanced capacity to identify vulnerabilities in critical systems.

The model, called Mythos, is not merely one more digital tool. It is a system designed to test the security of complex networks and infrastructures by simulating the behaviour of cyber attackers. Its function is to find flaws before criminals or adversaries do.

What alarmed

What surprised its own developers was the level of autonomy and creativity the system demonstrated. In internal tests, Mythos showed the capacity to explore unexpected paths, identify hidden gaps and persist repeatedly until it found a vulnerable point. One of those responsible for the project described the tool’s behaviour with a striking image: the artificial intelligence seemed capable of “scenting blood” — that is, of swiftly recognising the slightest signs of fragility and concentrating its efforts until it exploited the flaw.

Two sides

This metaphor does not signal literal violence. It expresses the extreme efficiency of a system that learns to detect vulnerabilities with a speed and precision surpassing human capabilities. In responsible hands, this technology can strengthen digital security. In hostile hands, it can facilitate attacks on an unprecedented scale.

It was this risk that led to the decision to suspend the model’s wide release. The fear was not merely technical but systemic: a tool capable of finding flaws so effectively could be used to compromise essential infrastructure, such as hospitals, banks, energy grids or transport systems.

The “phase of chaos”

It was in this context that the expression used by experts to describe the future scenario emerged: a “phase of chaos”.

This phase does not mean immediate, generalised disorder. It refers to a period of transition in which powerful technologies will be simultaneously in the hands of defenders and attackers, and in which the capacity to cause harm may grow faster than the institutional capacity to prevent it.

The turning point

The episode reveals a quiet shift in the way the world understands risk. For decades, innovation was treated as the undisputed engine of progress. Today, it has also become a source of systemic vulnerability. When a technology can compromise collective security, its governance ceases to be a technical matter and becomes a question of public interest.

The global agenda

This shift is taking place amid a reorientation of the international agenda. For many years, the fight against poverty and hunger occupied the centre of global priorities, consolidated in the Millennium Development Goals and subsequently in the 2030 Agenda. However, the intensification of armed conflicts, geopolitical instability and the advance of high-impact technologies have displaced the focus of international policy.

In 2024, the member states of the United Nations (UN) adopted three instruments that formalise this inflection: the Pact for the Future, the Global Digital Compact and the Declaration on Future Generations.

Security and ethics

These documents recognise that security — including digital security — has become an indispensable condition for sustainable development. This is not about abandoning the social agenda, but about recognising that public policies depend on institutional stability and technological predictability.

In this scenario, international concern is growing over the use of autonomous weapons systems, often called “killer robots”. These technologies were developed and tested before the construction of widely debated regulatory frameworks. Technical advance preceded ethical deliberation.

Maturity

The Mythos case follows the same logic. The technology became capable of identifying vulnerabilities on an unprecedented scale, and its dissemination could generate immediate risks to collective security. The decision to halt its release does not represent a scientific retreat. It represents institutional maturity.

Artificial intelligence has become critical infrastructure. Like energy, telecommunications and transport, it underpins the functioning of economic and social life. When a technology reaches this level of centrality, its governance must keep pace with its power.

Future generations

Among the documents adopted in 2024, the Declaration on Future Generations introduces a relevant principle: decisions taken today must consider their impact on people who have not yet been born. This perspective broadens the horizon of public policy and reinforces intergenerational responsibility.

The interdiction of Mythos can be interpreted as the first concrete gesture of this shift. It demonstrates that technological capability must be accompanied by mechanisms of collective responsibility and international cooperation.

The closing

The so-called “phase of chaos” is not inevitable. But avoiding it requires anticipatory institutional decisions, regulatory transparency and a commitment to public security.

The world has entered a new historical stage.

A stage in which innovation remains essential, but governance has become indispensable.