Artificial intelligence is not only forcing democracies to decide how to regulate a technology advancing at remarkable speed. It is raising a more uncomfortable question: whether the rules through which they protect individuals, data and freedoms could, over time, create an advantage for political systems that do not operate under the same constraints.
That is one of the most consequential points in an op-ed by Giorgia Meloni published in La Stampa, in which the Italian prime minister makes the case for global rules on AI while asking how democracies can protect what distinguishes them without undermining their ability to compete.
Why it matters: Meloni’s intervention comes as the debate around AI is moving rapidly from what increasingly capable models can do to the consequences of deploying them at scale.
- Tech companies are grappling with the implications of more powerful systems, while U.S. President Donald Trump is pushing the technological race forward with AI increasingly embedded in competition with China — a subject likely to feature when Xi Jinping visits the White House. In the background is the widening tension between power and safety, made more tangible both by hostile uses of AI models and by the growing autonomy of advanced systems.
In her words: Meloni starts from the individual. AI, in her account, can increasingly affect the production of knowledge, information, employment and, with the emergence of autonomous agents, decision-making itself.
- “Living in a world in which it is no longer possible to distinguish what is true from what is false means giving up the foundation of democracy,” she writes about synthetic content.
- On deepfakes, she adds: “We have a duty to defend personal dignity against degrading forms of manipulation.” Her proposed response includes binding global standards for tracing AI-generated content and systems capable of certifying digital identity.
The key shift: The political dimension runs through Meloni’s op-ed, but its international implications become clearest toward the end. She considers the argument that regulating AI could weaken the West’s current competitive advantage — then turns it around: how certain are we that this advantage will last?
- “If it is true, as it is, that AI is above all a multiplier of data,” Meloni asks, “are we sure that in the long run it will not be regimes that have the largest amount of data to process, since they do not have the rules, for example on privacy protection, that democratic systems have?”
From individual risk to strategic competition. The question changes the scale of the argument. The rules through which democracies protect privacy, accountability and individual freedoms also impose limits on the collection and use of data that authoritarian systems do not necessarily share.
- Those systems can also experiment with certain applications under fewer legal constraints and use state administrations and infrastructure more directly to support deployment.
- The issue, then, is no longer only how AI should be governed. It is whether differences between political systems could eventually affect the distribution of technological power.
The “overtake” is about the stack. Meloni’s warning points to a real asymmetry: AI competition is taking place between systems that do not impose the same constraints on technology.
- And an authoritarian “overtake” should not be understood simply as producing the world’s most advanced model. Power in AI will depend on the ability to combine models with data, compute, cloud infrastructure and applications — and then deploy that ecosystem at scale.
The big picture: U.S.-China competition already reflects this logic. The contest over advanced semiconductors, export controls, computing capacity and technological self-sufficiency is about control over an ecosystem rather than a single technology.
- The ability to spread that ecosystem matters alongside the performance of the models themselves.
The caveat: Data asymmetry does not predetermine the outcome. Compute, chips, energy, capital, talent and research remain decisive.
- The risk becomes more significant if fewer constraints on data use combine with industrial capacity, infrastructure and large-scale deployment, allowing an advantage in one part of the system to propagate across the stack.
- For democracies, the danger would then extend beyond losing technological leadership. They could find themselves operating in a global ecosystem whose rules have increasingly been determined elsewhere.
Why Meloni wants global rules. This helps explain Melo.ni’s insistence on the global dimension. If certain safeguards apply only to democracies, the asymmetry could widen. If rules on synthetic-content provenance, authentication, accountability and other safeguards become international standards, some of that gap can be reduced.
- But this is where the current paradox emerges. The need for shared rules is increasing precisely as AI acquires greater industrial and strategic value.
The logic: Washington and Beijing may have an interest in limited arrangements around specific shared risks. Yet competition over technology, infrastructure and standards makes comprehensive global AI governance considerably less plausible, at least in the short term.
- Meloni’s call could therefore translate more realistically into the gradual internationalization of rules around specific problems — synthetic-content provenance, authentication, serious incidents, responsibility for autonomous systems and some safety standards — that can eventually become sufficiently widespread to serve as market benchmarks.
Rules require power: Global standards do not prevail simply because they provide stronger safeguards. The economic and technological weight of those promoting them matters, as does the international reach of the systems incorporating those standards.
- This is where Meloni’s warning acquires its broader strategic significance. The democratic response to a potential authoritarian advantage cannot be to abandon the protections that distinguish democratic systems. But regulation alone will not be enough either.
The bottom line: If the West wants data protection, accountability, content authenticity and human control to become features of the global AI architecture, it will need sufficient technological and industrial capacity for those standards to spread alongside the technologies that embody them.
- Otherwise, democracies could remain highly effective at determining which rules apply within their own borders while systems built around different principles progressively acquire the power to determine which rules apply elsewhere.
