Una alerta global sacudió a PHP y dejó a la Argentina sin un hilo de respuesta oficial

El estero amaneció quieto, con esa luz pálida que suele recortar las copa de los árboles y dibujar el agua como una lámina plateada; en esa SAME early morning, while cybersecurity teams around the world were installing urgent patches, the local institutional response was notably absent. The image of the vulnerable web didn't register on social media, nor did it appear on the website of the National Cybersecurity Center, which has remained "under construction" since May. The absence contrasts with the coordinated warnings issued between Tuesday and Wednesday by the governments of the United States, the United Kingdom, Japan and several European countries.

The origin of the alert was an artificial intelligence model, GPT-5.6, which is no longer OpenAI's most recent version and contributed to identifying critical flaws in PHP. This programming language underpins approximately seven out of every ten websites worldwide. The vulnerabilities could have allowed malicious actors to access systems, extract private information, alter data or reach critical infrastructure, affecting municipalities, provincial governments, small business platforms and collaborative encyclopedias.

A global response in record time

Faced with the scale of the problem, OpenAI worked alongside researchers and the PHP team on an initiative called "Patching the Planet". Developers published security updates and technicians around the world moved to install them on affected servers as quickly as possible. According to those who participated, it was one of the largest and fastest coordinated response campaigns in recent cybersecurity history.

The official warnings from various governments urged system administrators to act immediately. Some even temporarily suspended public services while they completed the updates, a measure that shows the potential impact of leaving critical systems exposed. The speed of the campaigns also highlights how artificial intelligence tools can now trace vulnerabilities at a pace that far exceeds human capability alone.

• The flaws affected PHP, a language used by approximately seventy percent of the world's websites.
• The risks included unauthorized access, data theft, data modification and access to critical infrastructure.
• Authorities in the United States, the United Kingdom, Japan and several European countries issued coordinated alerts.
• Some governments temporarily suspended services while security updates were applied.
• The National Cybersecurity Center did not publish warnings on its social media accounts or website.

In Argentina, the institutional vacuum was even more evident. The National Cybersecurity Center, presented in May 2025, did not release a general guideline for handling the security issue. According to governmental sources, some entities acted on the initiative of their own technical teams, but there was no guarantee that all affected systems were protected. The center's website has displayed a construction notice since its launch, and the previous night its two main authorities resigned amid the crisis, just before the organization's first public event.

I have seen how a warning can become as important as the technical patch itself: without coordination, even prepared teams hesitate, and without clear institutional communication it's difficult for smaller organizations to know what to prioritize. The global campaign demonstrated that coordination is a form of prevention, and not merely an administrative step. In a country where the center created to watch these issues had no online presence and no published alert, the absence of guidance became the main news.

For those who manage a website, the concrete recommendation is to check the PHP version in use, apply the available security updates and verify that the changes have been correctly implemented, rather than assuming that a site is protected just because no incident has been reported. For citizens and small organizations, it's essential to contact their service providers and demand clear information about the patches applied, keeping in mind that the vulnerabilities could affect much more than a simple personal page.

The episode left a useful lesson: artificial intelligence can accelerate the detection of vulnerabilities, but it cannot replace the institutions responsible for communicating risks and coordinating action. The global response was rapid and coordinated; Argentina faced the same technical problem with a notable lack of official guidance. I'll be attentive to the publication of new warnings and the institutional situation of the center, and I invite readers to share any information about how their organizations handled this security alert.