The Central Election Commission decided to replace commission members with a neural network when checking subscription lists

The Russian Central Election Commission is testing the verification of subscription lists with a hardware-software complex involving a neural network, commission member Anton Lopatin told Interfax.

“The system is in trial operation, and the neural network model is now being trained to minimize the human factor when checking subscription lists and avoid speculation on this topic,” the agency quotes him as saying.

The complex should automatically check the data in the signature sheet with the electronic register of voters.

“If there are any differences in the data in the register, the expert will determine whether these could be features of the voter's handwriting or a falsified signature,” Lopatin said.

Now the compliance of the data in the signature sheet is checked manually.

• In elections in two regions of the Russian Federation, this method of verification has already been tested, but in parallel with the work of experts, Lopatin claims. He did not specify what level of elections and in what regions he is talking about.
• about automatic verification of signature sheets in February 2020 wrote Kommersant . How exactly the CEC plans to do this was not reported then.
• A large number of rejected voter signatures is a common reason for refusal to register for elections. For this reason, most independent candidates for the Moscow City Duma were not registered in 2019.