The new bill (if approved) will come into force on 1 September 2027 and is intended to apply to entities involved in the development, deployment and application of AI technologies in Russia. All foreign AI solutions somehow used in Russia may be impacted.
The bill will regulate the most critical AI-related issues, including IP rights, labeling and transparency, restrictions in the government sector and critical spheres, as well as liability and other areas.
Below is a high-level overview of the upcoming changes. Please, however, note that the bill is at the public discussion stage and may change significantly.
1. The AI Ecosystem: Key Roles & Responsibilities
The Bill defines the key actors in the AI value chain. Understanding your role is the first step to compliance.
2. Liability
General rule — all AI actors bear liability for harm caused by AI, proportionate to their fault.
Developers, operators, and service owners are liable for the result obtained using AI if they knew or should have known of the potential harm. To be exempt from liability, they must prove that they took exhaustive measures to prevent the incident.
3. Core Concepts of AI Models
| Concept: | What It Means for Your Business: |
| Trusted AI Model | For work with the government or critical infrastructure the AI model must be certified and listed in a state registry. All data processing in such model must be localized |
| Sovereign & National Model | AI models developed and trained in Russia by Russian entities using data sets formed in Russia by Russian entities will receive state support. The Government may establish cases where application of such models is mandatory |
***
We remain at your disposal and will be happy to analyze how the bill may impact your business model and develop a proactive compliance strategy.
Contacts
Alexander Monin, Partner
Ekaterina Abyasova, Senior Associate
Valeriya Eystrakh, Senior Associate
