Large language models: the new battlefield of Russian information warfare

The report warns that Russia has evolved its foreign information manipulation into a strategy of 'LLM grooming', flooding the web with low-quality disinformation (notably via the Pravda/Portal Kombat network) to be scraped by and influence large language models and AI chatbots; investigations by Viginum, NewsGuard, and university researchers show these narratives appearing in multiple chatbots and highlight the global risk to information integrity and democratic resilience.