China’s AI-Generated Viral Images Cause Division Among US Voters, Microsoft Warns
China has been utilizing generative AI since March in order to create viral content that aims to divide US voters on social media platforms, and according to Microsoft analysts, their efforts have only improved over time.
In a report released on Thursday by the Microsoft Threat Analysis Center, Clint Watts, the general manager, wrote that China-affiliated actors have been deploying AI-generated visual media as part of a broad campaign with a keen focus on politically divisive issues, such as gun violence, as well as denigrating US political figures and symbols.
Microsoft emphasized the potential impact of these efforts, warning that real individuals are unknowingly reposting these divisive images, mistaking them for genuine content. The quality of these AI-generated images is steadily improving, capturing the attention of authentic social media users, even though they exhibit traits commonly associated with AI-generation, such as hands with more than five fingers.
The analysts further noted that it has become increasingly challenging to differentiate between genuine and manipulated accounts. In the past, accounts linked to the Chinese Communist Party (CCP) would utilize computergenerated usernames, display names, and profile pictures that were easy to identify. However, the new operations are employing real individuals who adopt fictitious or stolen identities to mask their affiliation with the CCP.
While it may be difficult to detect these accounts, Microsoft outlined some tell-tale signs. Many of these operatives commence posting in Mandarin before subsequently switching to another language. Additionally, these accounts engage in a pattern of posting, commenting, liking, and sharing content from other similar accounts, creating an illusion of authenticity and boosting each other’s posts.
Microsoft analysts anticipate that China will continue to refine and develop this technology over time. However, the extent and timing of its deployment on a larger scale remain uncertain.
In response to Microsoft’s report, Liu Pengyu, a spokesperson for the Chinese Embassy in Washington, D.C., dismissed the accusations, stating that they were full of prejudice and malicious speculation against China.
This revelation comes after cybersecurity research firm, Mandiant, disclosed in July that Chinese-affiliated operatives had paid individuals in Washington, D.C., to protest a US government ban on goods originating from the Xinjiang region. Videos of these protests were then amplified by social media accounts associated with the operatives.
Furthermore, Meta (formerly Facebook) recently cautioned of a Chinese influence campaign that disseminated false information through 7,700 Facebook accounts, 950 Facebook pages, 15 Facebook groups, and 15 Instagram accounts. These accounts were subsequently removed by the social media company.
The continuous development of AI technology brings with it new challenges for social media platforms and governments to identify and counter disinformation campaigns. As the landscape evolves, it is crucial for individuals to exercise critical thinking skills and verify the credibility of the content they encounter on social media.
