The rapid development of artificial intelligence is creating not only technological breakthroughs but also widening the gap between countries and social groups. This is the conclusion reached by 40 independent scientists and experts who prepared a preliminary report for the UN Independent International Scientific Panel on AI.

The key problem, in my deep conviction, lies not in access to AI tools themselves, but in control over the infrastructure. Countries that import technology risk losing sovereignty over standards, security, and the adaptation of solutions to local realities. This creates a classic model of digital colonialism, where the supplier dictates terms and the consumer remains vulnerable.

Statistics confirm this thesis: the United States controls about 75% of the computing power of the world's leading AI supercomputers, while China accounts for approximately 15%. Together, these two countries accumulate about 90% of all computing used to train the most powerful systems. Such concentration inevitably leads to increased authoritarian control and undermines democratic accountability, as the report's authors rightly note.

Language Barrier and Critical Error

Language inequality deserves special attention. Generative models demonstrate high-quality performance with English and widely spoken languages, but for less common ones, such as Tigrinya (spoken by 7–9 million people), the results can be catastrophic. An example from the report is shocking: the machine translation system confused smallpox with syphilis, gonorrhea with diabetes, and interpreted a phrase about intravenous antibiotics as "intravenous insecticides." This is not just a mistake — it is a direct threat to life.

Basic internet access remains a fundamental barrier. According to the International Telecommunication Union, 2.2 billion people will remain offline in 2025. For these regions, AI implementation is not about algorithms but about electricity and digital literacy.

Recommendations and Risks

The report does not establish mandatory rules but offers clear directions for policymakers: developing local AI infrastructure and data centers, improving AI literacy, investing in developers, creating safety institutions, and strategies to combat disinformation. It also emphasizes the need to test AI systems after deployment on real users.

Risks of disinformation, fraud, election interference, and pressure on labor markets are highlighted separately. Let me remind you that in June, the Bank for International Settlements already warned: the investment boom around AI, which supported the global economy in 2025, is itself becoming a source of macro-financial risks.

Expert conclusion: The gap between technological capabilities and risk management mechanisms could indeed lead to catastrophic consequences. The cryptocurrency community, accustomed to decentralization, should pay particularly close attention to this trend — the concentration of AI power among a narrow circle of players threatens not only the global economy but also the fundamental principles of digital freedom.