AI in key spheres

The US experience

From November 2025, the Genesis¹ Mission was launched in the US to create an “American Science and Security Platform” for scientific data for the purpose of training fundamental models and “AI agents” that test hypotheses and automate research in priority areas: energy, scientific research and development, and defense. The platform was² developed as part of the “American AI Plan” (see review in Monitoring No. 7 (19) (July 2025)).

The platform should provide computing power, modeling and analysis tools, secure access to data, etc. There are plans to develop³ AI in the fields of advanced manufacturing, biotechnology, critical materials, nuclear energy, quantum computing, semiconductors, and microelectronics.

It is planned to develop model agreements on data and models, rules for licensing and commercializing intellectual property created using the platform, as well as uniform procedures for accessing data, models, and calculations, and cybersecurity standards for non-governmental participants.

Currently, 53 companies are planning to participate in the Genesis Mission in the fields of cloud technologies and general-purpose AI platforms (5 companies), hardware and data center infrastructure for AI (10), semiconductors, lithography, and critical materials for AI infrastructure (12), energy, including nuclear energy (6), quantum computing (2), and others. The Genesis initiative involves businesses in the development of AI solutions, including through budgetary funds. There are plans to involve startups.

The experience of China

In August 2025, opinions were published on deepening the implementation of the Artificial Intelligence+ initiative on the formation of programs, standards, and legal regulation of AI. The document contains the following directions:

- For science and technology – the development of scientific models and the creation of open and shared high-quality scientific data sets.

- Introducing AI into business strategies and processes and applying AI in design, production, and maintenance.

- Improving the system of data rights and copyrights as they apply to AI, opening access to data created within the framework of government-funded projects, and in parallel, developing national computing infrastructure (AI chips, supercomputing clusters, etc.).

- Establishing rules for monitoring AI systems, preventing risks, and responding to incidents, taking into account the level of risk. Here, China's approach to AI regulation is similar to that of the EU.

Unlike the US, China's approach is not built around a single specific state platform with⁴ a strict partner admission regime. Instead, China is implementing AI “in all sectors” of the economy and government, while simultaneously planning to “fine-tune” the framework rules: data rights and copyrights, standards, security assessment and registration systems, as well as monitoring and responding to AI risks.

Russia’s experience

In November 2025, Russia's Ministry of Digital Development, Communications and Mass Media discussed launching a pilot project on labeling AI-generated content, which would make it possible to distinguish between content created by humans and machines, as well as mechanisms for verifying the identity of users who post materials. It was proposed to analyze the current regulation of AI, assess the extent to which existing norms protect the rights of citizens when using AI in healthcare, education, and social support, and where gaps remain—to reconsider the basic norms of the law on AI.

In the future, laws on AI will be developed in Russia and around the world, defining the principles for using the technology and distributing responsibility for the correct operation of AI systems in critical areas (social services, healthcare, etc.) between the state, developers, and digital service providers. On average, in developed countries, about 15% of all AI projects will be funded by governments in the coming years. Today, this share ranges from 3% (US ) to about 50% (China ).

From the monitoring issue No. 11 (23), November 2025. Download the full issue (PDF) · issue page at the Gaidar Institute