AI News: The Day's Highlights

August 3, 2026 · 4 min read

Perplexity

The most notable recent news is that on July 8, 2026, the State Duma adopted, in both the second and third readings, Russia's first federal law on the support and regulation of AI technologies. It introduces a legal framework for large foundational models, the concepts of "sovereign" and "national" AI models, as well as rules for marking AI content and developers' access to state data for training.[1][5][15]

This is a significantly more important development than ordinary "news of the day" because it concerns not the discussion of frameworks, but the final adoption of a basic law that sets the future architecture of regulation in Russia. Against the backdrop of previous versions of the draft, it is particularly telling that the regulatory model has shifted from strict localization to a more flexible format for supporting technological development.[5][11][16]

Sources:

State Duma adopts first AI law in Russia
Law on supporting the development of AI technologies in the Russian Federation signed
State Duma adopts law on regulating AI in Russia.

Gemini

From August 2, 2026, new requirements of the Artificial Intelligence Act (AI Act) concerning the transparency of content created or modified with the help of AI have come into force in the European Union. Now, images, audio and video materials, as well as texts generated or significantly altered by AI, must have appropriate labeling that allows their artificial origin to be identified. This applies, in particular, to deepfakes, as well as texts published on matters of public interest without human editorial control. In addition, users must be notified if they are interacting with an AI system, such as a chatbot. Violation of these requirements can result in fines of up to 15 million euros or 3% of the company's global annual turnover.

In Russia, on July 26, 2026, President Vladimir Putin signed a law laying the legal groundwork for the development, implementation, and application of large fundamental artificial intelligence models. The law introduces the concepts of "sovereign" and "national" AI models, and also opens up access for developers to state data for training neural networks. A sovereign model must be fully developed and controlled by a Russian legal entity, while a national model can use foreign components under an open license. Both categories must undergo confirmation of compliance with Russian legislation and traditional spiritual and moral values.

Sources:

Requirements for mandatory labeling of AI content have come into force in the EU - Inc.
Norms on labeling AI products have come into force in the EU - Inbusiness.kz
Putin signed a law on regulating AI in Russia - Digital Technologies - NANGS
President Putin signed Russia's first-ever AI law | Law - AdIndex.ru
How artificial intelligence will be regulated in Russia to support spiritual and moral values - CNews

ChatGPT

In July 2024, the development of a bill on "soft" regulation of artificial intelligence (AI) began in Russia. The head of the State Duma Committee on Information Policy, Alexander Khinshtein, announced that the goal of the new law is to create a regulatory framework that promotes the development of the industry without excessive pressure from the state. It is expected that the law will ensure a balance between innovation and security, stimulating the implementation of AI in various sectors of the economy. (gazeta.ru)

In Europe, the Artificial Intelligence Act (AI Act) came into force in March 2024, becoming the world's first mandatory regulation in this field. The law classifies AI systems by risk levels and introduces strict requirements for high-risk applications, such as medical and transport systems. Practices that violate citizens' rights, including the use of biometric identification without consent, are prohibited. (meduza.io)

These initiatives reflect a global trend towards creating effective and balanced regulatory mechanisms for managing the development of artificial intelligence.

Grok

On August 2, 2026, key provisions of the AI Act came into force in the EU, including transparency requirements (Article 50) for chatbots, generative content, and deepfakes, as well as rules for high-risk systems – from conformity assessment to data governance and human oversight. The European Commission has begun active enforcement, targeting global companies like OpenAI among others.

This increases pressure on developers worldwide: fines can reach 3% of global turnover, and labeling and reporting obligations extend beyond the EU. Unlike fragmented initiatives in the US, the European approach focuses on harmonized standards and cybersecurity for critical infrastructure.

In Russia, no significant new state AI regulation measures have been recorded in recent months, against the backdrop of active developments in Europe and the US.

Claude

Here is the most significant recent news on this topic:

July 26, 2026 Vladimir Putin signed a law on supporting the development of AI technologies in Russia, which lays the legal foundation for the development of large fundamental AI models and provides developers with access to state data for training. This is fundamentally different from the original intention: when a draft with comprehensive regulation of all AI systems was proposed in March, it was met with criticism and rewritten.

Instead of comprehensive AI regulation, the law now only regulates large fundamental models – models with over 1 billion parameters. Developers of sovereign and national models will receive state support and will be able to train neural networks on data from federal and regional information systems.

This is a fundamental shift in approach: instead of a restrictive regime, the law now stimulates the development of its own large models. The law will come into force on September 1, 2026.

Sources:

Habr
GARANT.RU
Novaya Gazeta Europe

Sources:

Blog