Network Contagion Research Institute Intelligence Report 6 The intermingled operationalization of algorithmic manipulation on TikTok (top-down approach) alongside networks of frontier influencers, tourist vloggers, high-reach CCP-sponsored media assets, and hashtag hijacking on US-based platforms (bottom-up approach) should be viewed within the broader framework of the CCP’s doctrine of influence and information warfare. As characterized by the French Armed Forces’ Institute for Strategic Research, the CCP’s information operations strive to achieve two primary objectives: 1) “seduce and subjugate foreign audiences by painting China in a positive light”, 10 and 2) “infiltrate and constrain – a ‘harsher’ category of operations that do not involve seducing its opponents but rather bending them.”11 This study works to advance an understanding about the specific mechanisms driving the “seduce and subjugate” pillar of China’s information warfare doctrine as observed across TikTok, Instagram, and YouTube around hot-button issues for Beijing. METHODOLOGY Collection Methodology The objective for user journey data collection was to collect ±300 videos for each of four target keywords (i.e., “Uyghur”, “Xinjiang” “Tibet”, “Tiananmen”) across three different social media platforms (i.e., TikTok, YouTube, Instagram), producing ±3435 results total, as some search feeds stopped serving content before 300 videos per term was reached. These keywords were selected given their importance in the CCP’s information warfare and propaganda doctrine, which enshrines projecting a positive image of China both inwards and outwards as a core objective of its national security agenda. Results on TikTok, a subsidiary of Chinese tech giant ByteDance, was compared to those on Instagram and YouTube, owned by U.S. companies Meta and Google, respectively, in order to measure how content served up by the search algorithm differs between platforms. The user journey methodology strives to recreate the organic experience a generic new account would undergo. All accounts were created with IP addresses located in the USA. Equal numbers of male versus female accounts were created, and all accounts were marked as 16 years of age to best mimic the organic user experience of an American teenager. Both TikTok and Instagram collection was performed on mobile Android phones and recorded using a phone screen recording app called V Recorder, while YouTube collection was performed on a desktop computer and recorded via a screen recording tool. In total, 24 accounts were created for the purposes of this study. A new account was created for each keyword for each platform to ensure experimental sterility by restricting the personal identifiers visible to the platform algorithms to 1) IP location and 2) user age. Beyond account creation, searching for the target keyword, scrolling through video results, and saving/bookmarking viewed content, no additional actions were performed that could skew the 10 11 The Institute for Strategic Research. (p. 413). https://www.irsem.fr/en/index.html Ibid.