Chinese disinformation campaign on YouTube.
Australian researchers from the Australian Strategic Policy Institute (ASPI) have uncovered a significant disinformation campaign on YouTube orchestrated by a network called “Shadow Play.” Between October and November 2023, ASPI identified over 30 YouTube channels employing AI-generated audio, stock images, and consistent production techniques to spread disinformation on various topics.
These channels, collectively producing more than 4,500 videos, promoted narratives suggesting China’s superiority over the United States in a “technology war” and targeted specific companies like Huawei while downplaying achievements of US firms like Apple. The disinformation campaign, with a global reach, raised concerns due to its potential to covertly influence public opinion.
One notable instance involved four channels discussing the sale of Chinese-made buses to Nicaragua, each featuring similar content and edited versions of the same video to evade spam detection. The report highlighted shared thumbnails, keywords, and references among the channels.
ASPI named the network “Shadow Play” and emphasized its extensive impact, accumulating nearly 120 million views and 730,000 subscribers by the time of the report. The campaign’s ability to reach a vast audience globally raised alarms about its potential influence.
ASPI traced the origins of the campaign to a YouTube account sharing content from “Deepin Moments.” Through a structured search and analysis of metadata, the researchers identified patterns, leading them to conclude that the actor behind the videos is likely a native Mandarin speaker, familiar with the Chinese news cycle. ASPI believes the actor is potentially state-supported, aligning with China’s geopolitical goals.
Upon discovering the network, ASPI reported it to YouTube, resulting in the removal of accounts and videos due to terms of service violations related to misinformation.
In response to the findings, ASPI recommended increased transparency on social media platforms regarding the use of generative AI in content creation, urging investigations into such campaigns using in-house technical resources. They also called for closer monitoring by stock image providers and text-to-speech services to prevent misuse. ASPI suggested updates to terms of service to prohibit the use of such content for spreading disinformation.
Furthermore, ASPI proposed that the Five Eyes intelligence alliance share detailed information with policymakers, declassify investigations into similar operations, and consider public attribution of information operations to enhance societal resilience against such campaigns. The researchers anticipate that the identified network is likely more extensive than the reported 30 channels.