China's AI Influence Campaigns: A New Threat to US Democracy?

As the 2024 US presidential election approaches, new research reveals that China is ramping up its AI-driven influence campaigns, targeting American voters with divisive content.

Recent findings from Graphika have uncovered that Chinese state-linked operations are becoming increasingly aggressive in their efforts to influence the upcoming US 2024 election. These campaigns have been using fake profiles of US voters on social media platforms to discuss sensitive social issues, spreading divisive rhetoric ahead of the presidential election.

The influence operations (IOs) were found to have utilized AI-generated images of American voters, featuring lifelike avatars likely produced using a UK-based commercial AI video creation platform. Despite the convincing appearance of these avatars, the campaigns were primarily categorized as 'spamoflauge'—low-quality, spammy content that failed to gain significant traction. None of the reviewed videos received over 300 views, highlighting the challenges in producing effective political content.

Old Tactics, New Tools

The report also shows evidence of "coordinated amplification," where fake accounts reshared the same content across a network, amplifying misinformation. The content typically aimed to undermine US democracy and the political process, pushing debates on sensitive topics like the legitimacy of the 2020 election and anti-establishment messaging.

This is not the first time China has been implicated in campaigns to influence US citizens. However, the focus seems to be less on backing a specific candidate and more on dividing the American public and creating distrust in the US political system.

The use of AI in political influence campaigns is likely here to stay, and as the technology evolves, it will become increasingly difficult to detect. While most AI content creation platforms state that their services are not for political use, moderating its application remains a significant challenge.