Smart Engines and INTELLECT: Document Forgery in Russia Up 338%

The number of crimes registered by the Interior Ministry under Article 327 of the Russian Criminal Code (forging, manufacturing and selling forged documents) rose 338% in 2025, from 397 to 1 739 cases. That is the highest figure in the past seven years. The finding comes from Smart Engines, a developer of document recognition technology, and the law firm INTELLECT, in their fourth annual report on forged-document fraud. The minimum estimate for the illegal market in forged ID documents reached 2.6 billion rubles.

The arithmetic is simple. The count grew 4.4 times in one year, adding 1 342 registered cases. The authors call the situation a serious deterioration, and the numbers make that hard to dispute.

What the report recorded

The announcement on Habr shows that the study is built on official Interior Ministry statistics. The authors separately estimated the size of the black market for fake ID documents. The report has three blocks:

• 1 739 registered crimes under Article 327 of the Criminal Code in 2025, versus 397 a year earlier;
• +338% year over year, the highest value in seven years of observations;
• from 2.6 billion rubles: the minimum estimate of the forged ID market;
• the fourth edition of the series, so the data is collected over time rather than for a single year;
• an analysis of new schemes involving generative AI and of technologies for detecting forgeries.

Where generative AI comes in

In the announcement, the authors only mention "new schemes with generative AI." The details remain in the full version of the report, and it would be wrong to retell what the public part does not contain. Industry practice shows which scenarios come up most often: synthesizing photos for documents, editing scans and photos of passports, and swapping the image during remote identification. Specific shares and mechanics are best sought in the study itself.

Caveats on the numbers

The first caveat concerns the article of the code. Article 327 is broader than the fake ID problem and covers various types of documents. The growth cannot be attributed entirely to neural networks.

The second caveat: the Interior Ministry counts registered crimes. Such statistics reflect both a real rise in forgeries and how actively they are detected and prosecuted. When banks, carriers and online services introduce automated checks, the number of detected cases grows even if the number of attempts stays the same.

The third caveat: the report was prepared by a vendor that sells document verification solutions. That does not invalidate the Interior Ministry data, but the 2.6 billion ruble market estimate is given as a minimum, and the calculation methodology deserves a careful read.

What companies should do

For businesses that use remote identification (banks, microfinance organizations, telecom carriers, rental and delivery services), the takeaways are practical:

• Check not only the data in the fields but also signs of forgery in the image itself: security features, traces of a re-photographed screen, and editing.
• Verify the document against external databases and match the photo to a selfie with liveness detection.
• Store the original images and verification logs so that disputed cases can be reviewed and the rules retrained.
• Measure the share of missed forgeries and false rejections on your own traffic, not from figures in a vendor's presentation.
• Review the threat model at least once a year: image generation keeps getting cheaper, and yesterday's verification threshold may now be too low.

The full text of the report is published on Habr. If it includes a breakdown by document type and attack channel, that breakdown will show what share of the growth is linked to neural networks.