Проблемы института независимой антикоррупционной экспертизы и перспективы автоматизации

Общественные науки и современность. 2022. № 6. С. 89–105.

Porosenkov G., Резников И. А., Зюбанов К. А., Тухватуллин Т. А.

The article studies the problems of independent anti-corruption expertise which the experts accredited by the Ministry of Justice are facing in practice. Possible solutions to each of the highlighted problems were proposed, some of which contain usage of the automation tools and, in particular, artificial intelligence. In order to highlight and further analyze the problems, the conclusions of independent anti-corruption expertise, which are publicly available, were studied; 46 independent anti-corruption examinations were conducted, the artificial intelligence model CorrFinder, which allows to partially automate the procedure, was trained. The model uses logistic regression methods, Word2vec and TF-IDF.A number of problems are highlighted, such as: the problem of «denial policy» of public authorities in terms of review and recognition of the conclusions of independent experts; the uneven level of training of experts; imperfection of the adopted method of anti-corruption expertise, including a closed and limited list of corruption factors, as well as the inability to appeal in court against a legal act as illegal because of the presence of a corruption factor in it; low level of usability and transparency of regulation.gov. It is proposed: to create councils of independent experts on anti-corruption expertise under the Ministry of Justice to institutionalize their work; to introduce specialized training before accreditation and for the subsequent work of experts; to significantly update the methodology of the examination by creating an open list of corruption-causing factors and listing them in the Federal Law “On Anti-corruption expertise of regulatory legal acts and their projects»; to ensure transparency of the way of conclusion based on the results of an independent examination on the regulation portal.gov; to create a form for submission of opinion with the necessary fields to be filled in by an expert in order to reduce operational risk; to consolidate the resources of stakeholders of the anti-corruption expertise to train a model that allows partially automating the procedure.

Keywords: искусственный интеллект, автоматизация, коррупциогенные факторы, artificial intelligence, automation, independent anti-corruption expertise, methodology of anti-corruption expertise, independent experts, corruptive factors, независимая антикоррупционная экспертиза, методика антикоррупционной экспертизы, независимые эксперты