In a recent development, Microsoft researchers have unveiled what they believe to be a network of counterfeit social media accounts originating from China, purportedly aimed at manipulating US voters through the utilization of artificial intelligence (AI). The revelations have sparked controversy and drawn a response from the Chinese embassy in Washington.
The Chinese embassy spokesperson dismissed these allegations, stating that they are “full of prejudice and malicious speculation,” and reiterated China’s commitment to the responsible use of AI technology. Nevertheless, Microsoft’s research report presents compelling evidence of a suspected Chinese information operation that shares characteristics with prior activities attributed to an elite unit within China’s Ministry of Public Security, as noted by the US Department of Justice.
While the researchers did not explicitly name the affected social media platforms, screenshots in their report suggested that posts appeared on platforms resembling Facebook and Twitter, referred to as X in the report. This discovery adds to the mounting concerns regarding the influence of foreign entities on the US political landscape, particularly in the context of the upcoming 2024 presidential election.
Microsoft’s report provided some examples of recent activity but did not delve into the specifics of how they attributed these posts to China. A Microsoft spokesperson clarified that their researchers employed a multifaceted attribution model, relying on technical, behavioral, and contextual evidence.
The campaign reportedly began using generative AI technology around March 2023 to craft politically charged content in English, with the intention of imitating US voters. Generative AI is capable of producing various media, including images and text, from scratch. This approach has allowed the counterfeit content to be more visually appealing compared to previous efforts by Chinese state actors, which primarily relied on rudimentary digital drawings, stock photo collages, and manual graphic design.
One striking example highlighted in the report was an AI-generated image, attributed to a Chinese account, depicting the Statue of Liberty brandishing an assault rifle alongside the caption: “Everything is being thrown away. THE GODDESS OF VIOLENCE.” The identified accounts attempted to appear American by listing their public location within the United States, sharing American political slogans, and employing hashtags related to domestic political issues.
This revelation underscores the ongoing concerns about foreign interference in US elections, with the US government previously accusing Russia of meddling in the 2016 election through covert social media campaigns. Additionally, there have been warnings of similar efforts by China, Russia, and Iran aimed at influencing American voters.
In the broader global context, Chinese tech companies are actively participating in the AI race. Tencent recently introduced its “Hunyuan Aide” chatbot, which it claims possesses capabilities comparable to top US competitors like ChatGPT. This development further highlights the intensifying competition in the AI sector, driven by both technological advancements and regulatory changes in China, which aim to keep domestic developers competitive while maintaining strict control over online information.
