EU AI Act вступил в полную силу: штрафы до €35 млн для высокорисковых ИИ-систем

August 2, 2026, the EU AI Act — the most comprehensive artificial intelligence law in the world — entered into full force across the European Union. Companies using high-risk AI systems in the EU are now required to comply with regulatory requirements or pay fines of up to €35 million.

What requirements have entered into force?

From August 2, 2026, all companies with high-risk AI systems on the EU market must undergo full compliance. The law classifies AI systems in medicine, education, banking, personnel recruitment, and critical infrastructure management as high-risk.

Key requirements:

• Mandatory risk assessment before the system enters the EU market
• Documentation of training data, architecture, and algorithms
• Human oversight mechanisms and the ability to shut down the system
• Registration in the EU AI Office database
• Transparency: users must know they are interacting with AI

Fines will amount to up to €35 million or 7% of the company's global annual turnover — whichever is greater. A separate fine — up to €7.5 million or 1% of turnover — is provided for misleading the regulator.

How the law was introduced in stages

The EU AI Act was adopted by the European Parliament in March 2024 and officially published in July 2024. The entry into force of the requirements was spread over two years to give companies time to adapt their products.

From February 2, 2025, bans on the most dangerous practices came into force: citizen social scoring systems, manipulative AI with subliminal influence techniques, unauthorized biometric identification in public places in real time. From August 2025, rules for developers of powerful GPAI models came into effect — they were required to submit technical safety reports and assess systemic risks.

"The EU AI Act is the world's first law establishing legally binding requirements for AI.

This is a historic decision in favor of trustworthy and human-centric artificial intelligence," states the official materials of the European Commission.

What this means

The EU AI Act applies to all companies selling or using AI in the EU — regardless of the country of registration. OpenAI, Google, Microsoft, Asian and Russian developers are required to comply if their products are available to European users.

Analysts call this the "Brussels effect": the world's largest single market sets a standard that global companies often extend to all markets. This is exactly how the GDPR worked in 2018 — personal data protection requirements de facto became a global standard. The EU AI Act aspires to play a similar role in AI regulation.

Frequently Asked Questions

Which AI systems are considered high-risk?

High-risk AI systems include those used in medical diagnostics, educational assessment, credit scoring, personnel selection, and critical infrastructure management — power grids, water supply, transportation. The full list is set out in Annex III to the EU AI Act.

Which practices are completely prohibited?

Since February 2025, the following have been completely banned in the EU: citizen social scoring systems, AI with manipulative subliminal influence techniques, biometric identification in public places in real time (exceptions — strictly limited law enforcement operations). The fine for these violations is maximum: up to €35 million or 7% of turnover.