In the Novosibirsk region, archives are going to be "revived" with the help of artificial intelligence. The state archival service of the region plans to introduce a facial recognition system that will be able to find people in old photographs and videos.
The idea is simple, but it sounds almost like science fiction: a person uploads a photo of their relative, and the system automatically checks millions of archival images and videos. If matches are found, the user will be shown where exactly their ancestor might have appeared in historical footage.
According to the head of the state archival service of the Novosibirsk region, Konstantin Zakharov, the necessary technologies already exist - there will be no need to develop them from scratch. The system is planned to be integrated into the existing digital archival platform of the region, which already combines internal databases and processes of the archival service.
Preparation for launch is virtually complete. Applications for funding have now been submitted, and the implementation itself, according to preliminary estimates, may take several months.
In the future, they want to expand the capabilities of AI. For example, the neural network will be able to be used for automatic verification of digitized documents before publication - in order to quickly find personal data and decide whether it is possible to open materials for general access. Currently, this work is performed manually by archive staff, which takes a huge amount of time.
The scale of the task is really impressive. The archival service of the Novosibirsk region unites the state and 34 municipal archives. Their funds contain more than 3.9 million documents, the oldest of which date back to 1726.
Every year, specialists digitize about 4 million pages of archival materials, and by 2025, the total volume of digital documents has already exceeded 50 million images. At the same time, archives process more than 55 thousand requests from citizens every year - and the introduction of AI can seriously speed up the search for the necessary information.