Pooya Mokhtari – Cyber affairs expert

Artificial Intelligence; from General Technology to a Component of National Power

Artificial intelligence is no longer merely an economic technology or a tool for increasing productivity; rather, it depends on a set of infrastructural, computational, data, industrial, and human capacities that directly affect national power. In the US-China competition as well, superiority is not limited to the quality of language models or the number of advanced chips, and access to electricity, data centers, chips, cloud networks, capital, expert manpower, and the government’s capacity to regulate and enforce regulations also matters.

The Atlantic Council’s study on the competition between the two “AI blocs” of the US and China in Africa shows that model quality alone is not determinative, and infrastructure, financing, regulation, government capacity, and the ability to absorb technology can also be effective in the competition’s outcome. Another report from the same institution on US-Africa technological partnership also emphasizes that regulating AI rules, training elites, and creating institutional networks are part of the geopolitical competition over this technology. Consequently, soft power, standardization, and technology diplomacy have also become components of AI power.

China also pursues the use of AI technology within such a broad framework. Research by the Merics Institute shows that the development of energy infrastructure and clean technologies in China strengthens the electricity and industrial capacity needed for advanced technologies, including AI; therefore, AI competition cannot be separated from competition over energy, industry, and supply chains.

The US and China; Competition over the Technology Ecosystem

The recent dispute between Washington and Beijing over “model capability extraction” indicates that AI competition has entered a stage where even the outputs of models and the method of reproducing their capabilities have become a national security issue. In September 2026, US security officials accused several Chinese companies of extensive use of the “capability extraction” process to extract the capabilities of American models, but the Chinese side rejected the accusation and called it the politicization of a technical issue; therefore, this case should be considered part of the intensification of the technological conflict between the two powers, and its details should still be evaluated within the framework of the parties’ claims.

At the same time, Trivium China has reported that the two countries have turned to dialogue about AI on the eve of their leaders’ meeting. This trend indicates that technological competition does not necessarily lead to disconnection, and Washington and Beijing simultaneously need mechanisms to manage risks and prevent technological competition from turning into broader instability.

In the meantime, open models can also become an important tool of technological influence. ASPI’s analysis explains that China, by developing models with open weights and creating institutional mechanisms for their transfer to other countries, links the supply of technology with standardization, training, and capacity-building. This approach provides countries of the Global South with cheaper access to advanced models, but at the same time can create new dependency at the level of infrastructure, standards, and data.

Hybrid Warfare; the New Arena of Cognition and Information

One of the most sensitive areas of AI application is cognitive warfare and information operations. Media reports about the AI threat landscape review report prepared by Anthropic say that the company has announced that some government institutions of countries have used the Claude model to produce propaganda, design cognitive operations, and monitor opponents. These cases have been raised based on the company’s own assessment and should be considered merely as a company’s claim about how its systems are used.

The strategic importance of this issue is not limited to a specific case. AI can reduce the cost of producing targeted content, creating fake personas, analyzing audiences, mass translation, generating messages tailored to different groups, and monitoring the information environment. As a result, cognitive warfare can transform from limited and costly operations into a more scalable process.

The North Korea example also shows that AI can multiply existing capabilities in cyber warfare. ASPI argues that this technology does not necessarily create a completely new capability but can drastically reduce the cost and time of executing existing capabilities.

Military Acceleration; Operational Advantage or New Vulnerability?

The military application of AI faces a fundamental contradiction; the very speed that creates operational advantage can make control and evaluation more difficult. FPIF warns about the Pentagon’s acceleration in deploying AI tools, stating that the widespread use of commercial systems in military infrastructure can put security discipline, data control, and assurance mechanisms under pressure. According to this report, the use of generative systems in the US military structure is no longer limited to limited experiments and has expanded to areas such as planning, logistics, supply chain, and administrative affairs.

The issue is not just a model error. A published analysis on “repeatable error” in AI emphasizes that the main danger arises when a systematic error can be reproduced many times and on a large scale. From this perspective, AI security depends more on the quality of evaluation, data validation, and the design of control mechanisms than merely on model accuracy.

Eurasia Review also argues that the next arena of AI competition is “institutional speed”; that is, how quickly governments can test, evaluate, authorize, standardize, and operationalize technology without losing legal control and correction mechanisms; therefore, the strongest model does not necessarily create the fastest strategic advantage, and institutional capacity for safe technology absorption is also determinative.

AI Governance; Technology Control or Preserving Advantage?

The debate over AI regulation oscillates between two main views. The Hoover Institution warns about preemptive regulation, stating that strict and early regulations may limit innovation and competition; in contrast, the expansion of military, intelligence, and economic applications of AI has increased the necessity of creating safety and accountability rules.

At the global level, great powers have not yet reached complete agreement on the limits of these rules. The US seeks to maintain its technological advantage, China creates new standards and institutions for the development and transfer of its models, and Global South countries are also trying not to be mere consumers of technology. The experience of international cooperation in peaceful nuclear technology also shows that technical standards, export rules, political restrictions, and unequal access to technology can affect the development path of developing countries; an issue that is repeated in a different form in AI governance.

For Iran, the main message of these developments is that AI cannot be pursued merely at the level of using ready-made tools. In conditions of increasing technological competition, processing power, data, indigenous models, cybersecurity, human resources, electricity infrastructure, data centers, and legal mechanisms will all be part of national power. On the other hand, reliance on foreign models can create dependencies in data, standards, and infrastructure.

Ultimately, AI is becoming a “multi-purpose power technology”; a technology that links economy, security, intelligence, cyber warfare, cognitive operations, and technology diplomacy together; therefore, the US-China competition over AI is part of the broader redistribution of global power. For Global South countries, including Iran, the main issue is creating independent capacity to benefit from AI, reducing vulnerability to technological domination, and active participating in the rules that are being formed.