From time to time, I like to take an unconventional angle to study a subject, as is the case with fake interviews, for example.
A few weeks ago, I saw a post on my LinkedIn feed that said, “What if China made AI free tomorrow?” It was about fifteen lines long, simply written, a fictional scenario, but not uninteresting for all that.
Fiction? Yes, we are talking about something that, although possible, is extremely improbable, but the recent past teaches us that the impossible or improbable nature of an event is today quite relative.
In the end, it made me want to dig deeper and inaugurate a series of “dystopian chronicles” to explore improbable scenarios, knowing that reality sometimes does join fiction, if only on points of detail.
And while I’m at it, I might as well start with this idea of free Chinese AI.
- In 2025, China makes its AIs free to the world, displaying a discourse of inclusion but pursuing a strategy of technological and geopolitical influence.
- The countries of the Global South massively adopt these tools, attracted by their power and zero cost, while Europe hesitates between ethical principles and budgetary constraints.
- The United States suffers a major shock: its economic model collapses, causing a shift towards regulation inspired by Europe.
- In 2027, cognitive dependency sets in: Chinese AIs influence thinking, decisions and educational content in many countries.
- In 2030, a fragmented world emerges, between Chinese domination, a transatlantic revival and the rise of local initiatives to preserve intellectual plurality.
The turning point: Beijing announces free AI for all
The year is May 2025. As the whole planet debates the regulation of artificial intelligence against a backdrop of tension between Europe and the United States, a dispatch drops like a bombshell. At an unexpected press conference, President Xi Jinping has declared that Chinese AI models will be made available free of charge worldwide.
China will therefore make its latest generation AI accessible to everyone, businesses, governments and citizens, as part of a program called “AI for the People”.
The official discourse speaks of technological justice. Xi Jinping asserts that “progress cannot be reserved for a technological elite” and that “intelligence”, which has been nourished by human knowledge, is a product of humanity that cannot be monetized.
China thus intends to position itself as a technological benefactor of humanity, as opposed to Europeans and, above all, American giants who market their AIs via subscriptions, licenses, and expensive APIs.
The discourse, referring to inclusion, the common good, and solidarity, hits the mark and is particularly appealing to low- and middle-income countries that were left behind in this new El Dorado.
Of course, many analysts are not mistaken and are warning about a hidden agenda on the part of the Chinese.
For those, behind this benevolent narrative, the objective of China is to redistribute the cards of global digital power. Free access is not a philanthropic gesture but a tool of economic and ideological policy.
There is said to be a desire to weaken economically the current dominant players. The big American businesses (OpenAI, Anthropic, Google DeepMind) base their model on paid access to their AIs, and by making these services free, China is creating a technological dumping effect. It is pushing its competitors to review their pricing, damaging their profitability, and hoping to create a massive movement of users migrating to its platforms.
The initiative is also presented as an instrument of ideological influence. AIs are not neutral; they produce answers, summaries, translations, suggestions and therefore representations of reality. By massively distributing its models, China exports its cultural and ideological biases, its norms, its vision of the world and society. Some even speak of algorithmic diplomacy on a large scale
Finally, free access risks creating structural dependency: the states and businesses that adopt these AIs will find themselves captive to a software ecosystem. Once usage is established, resilience becomes low. Beijing will have control over future developments, the metadata collected, and the trajectories of innovation.
Six months later: mass adoption and Western panic
From the summer of 2025, the “AI for the People” program met with dazzling success. Chinese models were massively adopted in dozens of countries, often without public debate via bilateral agreements or technical choices motivated by budgetary constraints.
In Africa, Southeast Asia and Latin America, many governments are integrating Chinese AI into their administrations. Their motivation is simple: these tools are powerful, multilingual, easy to deploy and, above all, free.
Beijing accompanies its program with aggressive technical diplomacy: infrastructure donations, sending engineers, subsidized training, creation of AI innovation centers in Dakar or Bogota. In six months, China has become the default supplier of artificial intelligence throughout the “global South,” those countries characterized by a low human development index and GDP per capita, mostly located in the southern part of the emerging continents.
More surprisingly, the European Union is not insensitive to the Chinese initiative and although the movement is much slower, it is very real. Local administrations, SMEs and universities are starting to use free Chinese AIs for lack of affordable alternatives. Here again, the financial argument prevails.
But this is not without tension: with on the one hand the values of digital sovereignty, data protection and ethics and on the other the reality of costs, perceived efficiency and ease of access, values are crashing against economics and we are talking about a debate between ideologues and pragmatists. The European authorities are caught in a paradoxical situation: how can they defend their founding principles while being dependent on foreign solutions?
But it is in the United States that the shock is greatest. Despite the efforts of Europeans, this is where the world leaders in AI are and they are seeing their economic model collapse. Users are rushing to free solutions, investors are withdrawing and the tech sector is beginning to make massive layoffs. President Trump, beside himself, convenes a task force on “national security versus foreign AI” while the ruined Elon Musk falls from grace. The bursting of the AI bubble, improbable a few months earlier, is now a reality (AI heading for an economic dead end?).
It’s a bit like a “Pearl Harbor moment”. One fine morning, America wakes up stunned and realizes that it has lost at its own game, that of imposing on the world a business model that no one can compete with. The Chinese strategic state is in the process of reorienting the center of gravity of global cognitive production.
Two years later: cognitive dependence becomes widespread
In 2027, a growing share of the world’s intellectual activity passes through Chinese AI models. The phenomenon is no longer technical: it becomes cultural, normative, cognitive.
AI is no longer content to answer questions, it organizes ways of thinking. It becomes the filter through which teachers, professionals and citizens interact with knowledge. Filters produced by entities aligned with Chinese geopolitical interests.
There are undeniable shifts. Quotations, examples, answers and figures of speech resonate with the party line.
Many governments are starting to sound the alarm about what they call the loss of cognitive autonomy, but the damage is already done. Brazil, India and South Africa have identified cases where Chinese AI tools are influencing public decisions, educational content and political priorities, and researchers are talking about “cognitive colonization”.
The question is no longer who provides the technology, but who controls the way of reasoning.
But it is in the United States that the most astonishing turnaround is taking place, with a major doctrinal change.
Until now, they have been espousing an ultra-liberal vision of innovation: letting businesses innovate without restriction, capturing market share, creating value. This model has been blown apart by a power that mobilizes the resources of a state to subsidize, structure and deploy at a loss.
Washington is therefore forced to reevaluate its doctrine: without regulation and public support, its economy cannot resist.
From 2026, the Trump administration begins to draw inspiration from the European approach: regulation of models, independent audits, algorithmic risk rating, transparency of data sets, public funding for open source models. A federal AI oversight agency is created.
This convergence gives rise to a new transatlantic alliance for a regulated, pluralistic AI we can trust, opposed to the Chinese logic of cognitive encirclement.
2030: tensions, alternatives and reinvention
In 2030, the world is more fragmented than ever.
The Chinese cognitive empire and its free AIs still dominate a large part of the uses in the Global South, where they have become critical infrastructures: it is impossible to deactivate them without collapsing educational, administrative and health systems. Their evolution remains opaque, the conditions of use are regularly modified, the dependencies are accentuated but these countries are trapped. Getting out of the Chinese ecosystem is becoming an industrial project in its own right, out of the reach of most of the governments concerned.
For its part, and after long procrastination against a backdrop of internal dissension, the European Union has engaged in an effort to regain technological sovereignty. Several Member States, led by France and the Scandinavian countries, have launched national or regional programs for the development of ethical, regulated and interoperable AI based on the once controversial framework of the AI Act (The European AI Act for Dummies).
They have structured an ecosystem based on open source, audited and sometimes shared models. These AIs, less powerful but more governable, have been integrated into administration, justice and higher education.
But this has come at a price: that of a two-speed Europe on the brink of division and only held together by the common defense project. Indeed, certain countries facing budgetary constraints or having made different political choices (Italy, Hungary, Slovakia, Portugal, etc.) have maintained or reinforced their use of Chinese models, particularly for local authorities, translation tools or social services. This has created a rift within Europe on a technological and strategic level, which is difficult to resolve and which blocks any form of cooperation when data exchanges are at stake.
Europe is therefore continuing its history by repeating its past: on the one hand, the consolidation of its normative and regulatory capacity, now internationally acclaimed as a perfect balance between innovation and security, and on the other, the confrontation with its own internal limitations, which still prevent a coherent and coordinated global response.
As for the United States, it has “Europeanized”. The free nature of Chinese AI had been perceived from the outset as a hostile influence operation and, as we have seen, caused a doctrinal shift with the renunciation of their tradition of non-intervention in digital innovation.
A series of federal laws regulated the production, dissemination and use of AI on American soil. Massive funding was directed towards universities, public agencies and industrial consortia to revive a technological offer aligned with democratic values, national security and transparency of models.
This turning point has resulted in an unprecedented convergence with Europe. Transatlantic cooperation has been established to pool ethical standards, create compatible infrastructures, and support the production of open, reliable, explainable models. Regulation, long perceived as a brake on innovation, has become a lever of geopolitical influenceand the foundation of a resurgent transatlantic dynamic.
But the United States has emerged weakened from this episode: its private technological fabric, weakened by gratuitous competition, struggles to reinvent itself in a more structured framework to which it is not culturally accustomed. A new generation of hybrid businesses, combining commercial logic and public service missions, is slowly emerging, driven by a new ideology that accompanies the renewal of Silicon Valley.
China, for its part, has succeeded in imposing its models as implicit standards in more than 70 countries. It has consolidated a network of sustainable technological alliances, particularly with middle-income states, regional organizations and foreign public companies.
But this strategic success also has its price. The country is indeed prey to growing internal weaknesses. The strict ideological control imposed on Chinese AIs is beginning to hamper their capacity for innovation. Technological entrepreneurs are expressing, often in veiled terms, their frustration at the excessive centralization of technological choices. Tensions are arising between the need for local adaptation of AIs (for example, for Africa or the Arab world) and the constraints of ideological conformity imposed from Beijing.
Externally, several governments have expressed concerns about the capture of sensitive data, the opacity of software updates, and the implicit standardization of generated content. Local initiatives to reclaim the models are beginning to emerge, with limited but symbolically important success.
Indeed, this context of extreme polarization has allowed a third space to develop: that of community-based, territorial, shared AI models. Cities (Barcelona, Montevideo, Montreal), universities, NGO coalitions and media are investing in the creation of shared governance AI models.
Their ambition is not pure performance, but collective control, transparency of processing logic, and the cultural anchoring of the responses generated. These AIs are limited in power, but aim to be rich in diversity. They give users back the power of arbitration and recontextualization and embody a form of distributed cognitive resistance, based on the relocation of uses.
They do not yet constitute a global alternative and are not intended to be one: they seek above all to preserve the plurality of thought and to avoid cognitive standardization on a global scale. But they demonstrate that another trajectory is possible if we think of AI as a common good to be governed, and not as a commodity to be consumed.
Bottom line
I often say that in AI, the technology is ultimately secondary to its scope and that, moreover, it is not so much a matter of companies as of countries confronting each other on different dimensions through companies that convey their ideology.
In this specific case, it is not a technology that has been exported, but an architecture of dependence: economic, political, cognitive.
Behind access to AI lies the control of the conditions of access to knowledge, of decision-making, of the formulation of thought.
The probability of a country imposing itself at the global level through dumping is more than low, but this fictional scenario shows us what we have to lose if a dominant player ignores all ethics and approaches that are criticized today, sometimes rightly so, can be our lifeline tomorrow.
Illustration: generated by AI (DALL·E – OpenAI)
