Spiritual screening for AI: Russia introduces mandatory checks of models for traditional values
The Russian artificial intelligence market is on the brink of tectonic shifts. This is not just about technical regulation, but about introducing a "spiritual and moral review" for AI models vying for the status of sovereign or national. This initiative, actively discussed in late July at the ANO "Digital Economy" platform, radically changes the rules of the game for developers seeking access to government contracts and preferences.
How will the new verification system work?
The AI law that came into force on July 26 is framework in nature, so the key mechanisms will need to be spelled out in bylaws. A developer seeking a high status for their model will have to prove its compliance not only with legislation, but also with certain "traditional values." Notably, no separate document listing these values is planned — it is assumed they are already enshrined in Presidential Decree No. 809.
Instead of a formal list, regulators, together with businesses, will develop special benchmarks — sets of tests to verify models. The procedure will be two-stage: first, the developer independently tests the model using a general risk-based methodology, and then an accredited laboratory evaluates these findings and may selectively check the system's resilience, including attempts to hack it via prompt injections.
The final conclusion on compliance will be sent to the government commission on digitalization. It is the positive verdict that will become the "golden ticket" to state support and advantages in government procurement. At the same time, as emphasized, not all models on the market will be checked — only those claiming strategic status. The first areas of implementation will be education and public services.
Who will become the arbiter on matters of spirituality?
The key question is who will issue this verdict. The scheme involves creating a pool of accredited laboratories, rather than a single designated center. Requirements for them will be developed jointly with businesses. The final decision on status will remain with the Ministry of Digital Development, while the FSB and FSTEC will check security for state systems.
Justified concerns are already being voiced in the professional community. Alexey Borshchov, business architect of AI projects in the Avandok ecosystem, rightly notes that categories such as "civic consciousness" or "historical memory" are difficult to translate into a technical specification. Meanwhile, Dmitry Galantsev, managing partner of the law firm Propositum, warns of the risk of "subjectivity" in the criteria, which could lead to selective application.
My analysis: This is a classic case where regulatory logic outpaces technological reality. Formalizing "spirituality" in the form of benchmarks is a non-trivial task, fraught with either a formal approach or the creation of a tool for covert control. For the market, this means rising costs and stronger entry barriers, which in the long term could slow down innovation. However, for major players who already have content filtering tools, this is a chance to strengthen their positions.