Russia's AI-generated disinformation campaign on Meta platforms
Russia conducted coordinated inauthentic behavior campaigns using generative AI to create fake journalist personas and fictitious news websites aimed at influencing the US election and rallying support for its war in Ukraine, according to Meta's security report. Meta said the AI-generated tactics provided minimal disruption and its detection systems remained effective against the operations.
What happened
Meta's quarterly Adversarial Threat Report, published 15 August 2024, identified Russia as the 'number one source' of coordinated inauthentic behavior with at least 39 covert influence operations. Russian operators used generative AI to create fake journalist personas and publish stories on fictitious news websites with summaries generated from authentic articles to make them appear more unique, along with AI-generated news-reader videos.
Reported impact
Meta said the Russian campaign focused on Russia's war in Ukraine, attempting to rally support for the invasion and influence the US election scheduled for November 2024. The report stated Russia was expected to promote candidates opposing aid to Ukraine and criticize those advocating Ukrainian defence, potentially by blaming US economic hardship on Ukraine aid, portraying Ukraine's government as unreliable, or amplifying pro-Russia views.
Index analysis
Meta reported that GenAI-powered tactics provided 'only incremental productivity and content-generation gains' and had not impeded its ability to disrupt the operations. The company described the networks as 'low-quality, high-volume' with security lapses, noting that real people frequently identified these networks as trolls and they struggled to engage authentic audiences.
Governance controls
Meta said it targets and removes deceptive posts and accounts relying heavily on AI or operated by contractors in for-hire deception campaigns. The European Commission is conducting a probe into Meta's Facebook and Instagram handling of Russian disinformation under the Digital Services Act, though findings remain pending.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.