“The thing secretly feared most always happens. I write: Oh You, have mercy. And then?” wrote the Italian writer, Cesare Pavese, on August 18, 1950, in his diary The Business of Living (Il mestiere di vivere).
Well, even within the mundane digital landscape, the most feared outcome has come to pass: an authoritative intervention by the US government has blocked Anthropic’s two flagship Artificial Intelligence models, Fable 5 and Mythos 5.
A Friday afternoon, 5:20 PM in New York. A single-line order from Washington. Two AI models—the most advanced in existence—vanish from the screens of users, corporations, and researchers in every corner of the globe.
Claude Fable 5 and Claude Mythos 5, Anthropic’s technological crown jewels launched just three days prior, abruptly go dark.
This is neither a technical glitch nor a cyberattack. It is the raw power of state sovereignty bursting into Silicon Valley’s gilded garden.
The United States government has invoked national security powers and export controls.
The reason? To block access to the models for any foreign national, whether inside or outside American borders—including Anthropic’s own employees who do not hold a US passport.
The company’s response is an effective total surrender, accompanied by a statement of profound political dissent.
To ensure compliance with the executive decree, Anthropic has chosen to shut down international access for all customers.
It is a radical decision that tears away the veil of technological neutrality, exposing the fragility of the digital infrastructure upon which we are building the future of labor, law, and knowledge.
Navigating this scenario requires looking past the noise of breaking news.
Anyone seeking conventional frameworks to understand the present knows that this event is no mere speed bump.
It is the exact intersection where media studies theories transform into geopolitical reality—an open clash between state power and the presumed autonomy of digital giants.
The guardian betrayed by its own prayers
The situation takes on an ironic twist when examining Anthropic’s trajectory. We are talking about a laboratory that built its public identity on the concept of AI safety, caution, and oversight.
For months, the company’s leadership called for public intervention, government testing, and even the state’s power to shut down systems deemed dangerous.
They made these requests within a specific framework: clear rules, transparent procedures, and objective technical evaluations.
Instead, US political power responded with an ancient, blunt instrument: export controls on national security grounds. No debate, no transparency.
Only an executive order based on a hypothetical vulnerability—a “jailbreak” method that, according to Anthropic, merely allowed for the detection of already known software flaws.
These are flaws that other widely used models, such as OpenAI’s GPT-5.5, manage and expose every day without facing censorship.
This short-circuit recalls the insights of media sociologist Denis McQuail in his foundational work on the relationship between journalism and society.
McQuail reminds us that the Internet was born without preassigned purposes or rigid regulations, only to later evolve into a cluster of large, profit-driven private enterprises.
Today, that potential for an open civic space is shrinking.
Digital systems are becoming targets of state surveillance, interference, and legal incursions.
Anthropic invoked the law as an ethical shield, but instead found the law operating as the arm of power geopolitics.
The geopolitics of the “Platform Society”
In their essay The Platform Society, José van Dijck, Thomas Poell, and Martijn de Waal analyze the deep mechanisms governing the ecosystem of digital platforms.
The authors explain that Western tech giants do not operate in a vacuum; rather, they are deeply rooted in the American ideological and political system.
The neoliberal model has allowed the five tech titans—Amazon, Apple, Meta, Alphabet, and Microsoft—to accumulate a market value exceeding the gross domestic product of entire nations.
The CEOs of these Big Tech firms act as genuine political leaders—international figures who have already expanded their reach into the cosmos through space missions and satellites.
The US government’s decision to strike Fable 5 and Mythos 5 proves that this autonomy has an insurmountable limit.
When technological supremacy becomes intertwined with state security, the Washington administration does not hesitate to assert its control.
This is reminiscent, after all, of how absolute monarchies treated printing and publishing back in the 18th century.
The US government treats software code as war materiel, weapons, or sensitive aerospace technologies.
Van Dijck and his colleagues remind us that the platform ecosystem is anything but neutral.
A single company’s attempt to alter markets or scale its services translates into an effort to shift the balance of power.
With the June 12, 2026 order from the US government, the state reaffirms its supremacy over multinational corporations.
The global connectivity infrastructure is bent to the defensive needs of its nation of origin.
The intercultural ramifications of this shutdown are massive.
Banning access for foreign nationals, including the company’s own internal employees, means drawing an ethnic and national border directly inside the code.
It is the antithesis of the very concept of a global digital community.
Intercultural journalism compels us to examine the implications of this choice.
Here is a question that deserves to be asked: who decides which citizens are trustworthy and which pose a threat simply by virtue of being born elsewhere?
Technology, which was meant to unite us and transcend barriers, becomes an instrument of division and exclusion.
It feels like being thrown back to 1998, when I first began analyzing how the Italian media represented migrants and citizens of foreign origin.
During the hot summer of 1998, marked by the arrivals of migrants, the equation “foreigner equals criminal equals danger” emerged.
Today, the distinction drawn by the US government—which Anthropic found technically impossible to implement selectively—points in that exact direction.
It is the same xenophobic direction, moreover, that the Trump administration (headed by a man of German descent, no less) has been pursuing for months.
Data, culture, and the architecture of power
To understand why an advanced language model arouses such fear from a government, we must return to the categories outlined by Nicholas Carah and Sungyong Ahn in their essay Media & Society.
The authors indicate that media power today is built upon two interconnected social processes: representation and automation.
Representation is the ability to produce and distribute meanings, narratives, and symbols through which people interpret the world.
Automation is the process of collecting and processing data to delegate decisions to machines.
Models like Fable 5 and Mythos 5 merge these two aspects seamlessly.
They do not merely compute data; they generate meaning and write code.
They assist the judiciary, politics, and culture.
They manage communication flows and analyze critical infrastructure.
Controlling these models means controlling the contemporary factory of meaning and decision-making.
When the US government orders a halt due to an alleged flaw that allows the model to patch programming bugs, it is protecting its monopoly over this automation capability.
Whoever possesses the most sophisticated algorithmic engine holds an immense strategic advantage in managing power within fragmented societies.
Allowing foreign citizens access to this resource means, in the eyes of the Pentagon or the White House, surrendering a share of algorithmic sovereignty.
Anthropic’s choice to comply and disable the models for everyone, rather than attempting a complex technical segregation of access, exposes the intrinsic nature of platform capitalism.
As media studies experts note, platforms are engineering firms that translate social life into data.
Faced with the threat of government sanctions or a total revocation of operating licenses, the logic of profit and corporate survival dictates absolute compliance—at the expense of the service provided to millions of users worldwide.
The risk for citizens and professionals
The shutdown of Anthropic’s new AI models delivers a harsh lesson that shakes the foundations of the professional world.
Those who work with data, legal consulting, code, or information management discovered overnight just how fragile the ground beneath their feet truly is.
We have delegated substantial portions of our workflows to private architectures located overseas.
If a model can vanish in a matter of minutes due to a geopolitical whim, the issue directly threatens the very continuity of our operations.
This is the operational risk of digital dependence.
As a community of readers and professionals, we cannot afford blind faith in Silicon Valley’s tools. It is imperative to develop strategies for autonomy:
Keeping local backups of valuable materials.
Implementing redundancy plans utilizing open-source models.
Cleanly decoupling the algorithmic tool from our personal intellectual methodology.
Access to an application, no matter how extraordinary, is no substitute for a long-term strategy.
The infrastructures we once considered stable and universal stand revealed instead as precarious constructs—hostages to profound conflicts between blocs of nations.
Toward a new media ecology
Analyzing this clash between political power and the AI enterprise forces a reconsideration of the entire structure of the information society.
Media studies teach us that every technology carries an implicit ideology.
Platforms have long promoted a narrative of freedom, efficiency, and personalized services, masking the centralized concentration of control.
Now that the state is reclaiming its role as vigilante and guardian of digital borders, that original promise evaporates.
What remains is a complex web of relationships where we, the people, risk becoming the injured party—deprived of our work tools from one moment to the next with no right of appeal.
For working professionals and anyone who rejects simple answers, the Anthropic affair points in a clear direction.
The time has come to demand a public debate on the values that should steer technology.
We must shift our focus from individual applications to the broader multimedia ecosystem.
We need to demand safeguards, transparency, and forms of digital sovereignty that protect the common good from the twin grips of state power and corporate interests.
Only by cultivating a critical perspective can we navigate the complexity of our time without becoming its prisoners.
This is, after all, the lesson I have learned in over 45 years of journalism and nearly 30 years of media studies.
It is vital to foster critical thinking, sharpen our analysis of the events around us, and ground ourselves in an ethics of respect for individuals, academic freedom, and the free expression of ideas.
Maurizio F. Corte
- Maurizio F. Corte, a professional journalist, media writer, and media educator, is an adjunct professor of Intercultural Communication in the Media at the Center for Intercultural Studies of the University of Verona. He is also the academic coordinator for the Master’s programs in Intercultural Competence and Management, as well as European Journalism, Media, and Intercultural Journalism.
- Let’s stay in touch on LinkedIn.
(In researching and editing this text, the author utilized the AI models Gemini and NotebookLM).
Bibliographic Sources
Articles do not emerge from a vacuum. They are built upon high-profile readings alongside professional and personal experiences. Below are the primary sources the author drew upon.
This analytical work was supported by the ethical and responsible use of Artificial Intelligence, trained by the author for these articles. The AI models used were: NotebookLM, Claude, and Manus.
- Denis McQuail, Mark Deuze, Media & Mass Communication Theory, Sage, London, 2020.
Nicholas Carah, Amy Dobson, Sungyong Ahn, Media & Society, SAGE Publications, Londra, 2024.
Nick Couldry, Andreas Hepp, The Mediated Construction of Reality, Polity Press, Cambridge, 2016.
Leighton Evans, Media Studies: Industries, Texts and Audiences, SAGE Publications, Londra, 2023.
José van Dijck, Thomas Poell, Martijn de Waal, The Platform Society, Oxford University Press, New York, 2018.
Nick Couldry, Andreas Hepp, Wolfgang Reißmann, Media Practice Theory, Routledge, New York, 2023.
Deana A. Rohlinger, Sarah Sobieraj, The Oxford Handbook of Digital Media Sociology, Oxford University Press, Oxford, 2022.
Michael Stevenson, Misha Kavka, Doing Media Research, SAGE Publications, Londra, 2023.
Stanley J. Baran, Dennis K. Davis, Mass Communication Theory (8th Edition), Oxford University Press, New York, 2020.