An AI model created by MIM Solutions will effectively fight anti-Ukrainian content on Twitter.

Experts from MIM Solutions have created an AI-based model that independently detects anti-Ukrainian content on Twitter. The goal of the algorithms is to find Polish-language tweets and comments published as part of the information war waged by supporters of Russia. In doing so, the system identifies accounts that generate pro-Russian propaganda.

MIM Solutions is implementing and disseminating its PRO BONO project under the hashtag #AIForUkraine.

“As part of the brutal war in Ukraine, Russia is waging a parallel, aggressive, wide-ranging cyber war. We believe that all possible tools must be used to detect these propaganda activities. At MIM Solutions, we create NLP (Natural Language Processing) solutions based on deep neural networks that effectively flag such harmful content. We hope that by sharing the accounts found by our monitoring system it will be possible to more effectively combat propaganda activities. We encourage everyone to follow our @AIForUkraine Twitter account and assist us in responding to inappropriate content” said Piotr Sankowski, Chief Scientific Officer at MIM Solutions.

The aim of the “Algorithms for Ukraine” system is to automatically identify Twitter accounts that publish Polish-language content of a clearly anti-Ukrainian nature and to log them on a dedicated @AIForUkraine account.

The automatic search for hostile accounts is done by monitoring tweets and rating them. The accounts themselves are selected based on the overall information they provide. The efficiency level of the model is very high, over 95 percent.

You can read more about our project on the Dziennik Gazeta Prawna website in the article “Artificial intelligence will fight disinformation on Twitter”.

You can read the entire article (in Polish) here: Dziennik Gazeta Prawna.