This section of the report explores qualitative and quantitative evidence illustrating the global implementation of RNR policies as access barriers. In compiling this evidence, we selected a representative group of China’s most popular “Social” apps, which represents the “grey zone internet” as conceived by Xi Jinping in 2013. These apps, if freely available, would facilitate broad and meaningful information exchange between domestic and global users.

Whether or how app providers ought to apply RNR requirements to non-PRC based users, however, has never been explicitly regulated in relevant documents. Yet by the same token, overseas users are also not exempt from China’s expanding RNR system. This, in theory (or perhaps by design), excludes the use of overseas phone numbers as an RNR proxy for most users. Accordingly, overseas users on Chinese platforms exist in a legislative and political grey zone in which the CCP has operationalized identifiability to enable surveillance and transnational repression.

Given our research and findings, this paper argues that identifiability obligations on China’s social media platforms were designed (and in practice operate) as a protective membrane which can be strategically used to restrict the free exchange of ideas and to separate the Chinese internet from global online communities. Such actions turn RNR into a simple, yet effective, exclusionary mechanism applied to ensure domestic discourse security. Unsurprisingly, as discussed below, new access barriers have subsequently emerged around Chinese social media apps, especially in regions crucial for the exchange of (political) ideas with the Chinese internet.

Four types of barriers to keep out overseas users

Article 24 of 2017’s Cybersecurity Law required Chinese internet service providers, including social media platforms, to verify RNR before providing users with online services. This requirement forced Chinese internet companies to ramp up their RNR verification processes and account management practices, which ultimately posed significant registration challenges for overseas users. While RNR for Chinese users was achieved by linking phone numbers to accounts, internet companies had to develop new methods to ensure RNR for overseas users whose phone numbers could not be matched with identification documents held by Chinese telecom providers (see Figure 3, above, for more details).

The net effect of these regulations was to significantly increase the administrative and technical burdenfor companies supporting users outside of the PRC.¹ As a result, certain online payment providers (like Alipay) and mobile game developers (like Tencent) temporarily disallowed new foreign accounts.² In turn, the majority of Chinese social media platforms announced that overseas users would now be required to submit ID scans. ³ Video platform Kuaishou, however, pulled its app from app stores altogether due to the new administrative challenges.⁴

Commercial decisions like these highlight how Chinese companies have generally succumbed to pressure from their government, which demands that party interests be prioritized over operational or business interests.⁵ Studies have shown that the CCP’s extensive influence on tech companies extends deep into their products, including search-engine level censorship ⁶ and input-level censorship.⁷ Even U.S. tech giant Apple has removed numerous apps from its app store as a result of such pressure.⁸

While some companies temporarily halted new foreign accounts to comply with RNR requirements, others attempted to navigate the regulations by developing separate apps for different regions or implementing varying levels of access restrictions (see Figure 4, below). Strategies like these underscore the complex balancing act that Chinese companies must perform to align with the CCP’s policies while still pursuing international growth.⁹

Figure 4: Strategies adopted by Chinese app developers

Chinese app developers have adopted a variety of approaches to reconcile their global growth targets with the complex regulatory environment related to RNR. Research for this report identified three such strategies:

Different apps for different regions:

Five different versions of the short video app Kuaishou are available for different geographies. The original (快手) is available in China and many other countries, except in the MENA region and most South American countries.¹⁰ A “lite” version is available in China only (快手极速版).¹¹ There is also a version available only in the MENA region (Kwai - download & share video), ¹² a version only available in Latin American countries (Kwai - ver vídeos bacanas), ¹³ and a version for niche content available everywhere except Brazil and Kazakhstan (噗叽).¹⁴

One app, two systems:

Certain apps have split into two entities to cater to domestic and global audiences separately. Some allow limited exchange between users, such as Weibo Intl.¹⁵ and Weibo (微博).¹⁶ Others do not allow any interaction between global and domestic users — ultimately functioning as two separate online spheres, as seen with DingDing and DingTalk¹⁷, or TikTok and Douyin.¹⁸

Stay available under one ID:

Apps like WeChat¹⁹ maintain almost identical software packages across international app stores. Various local adjustments can be made, however, such as log-in requirements, default language, and personalized recommendations. This strategy offers less flexibility than the multi-app approach but allows for a more unified presence.²⁰

We identified four distinct types of access barriers that can exclude users or force identifiability during our app walkthroughs (from downloading software on app stores to engaging with other users on the platform). These included:

Type #1. App store censorship:

Apps are pulled from app stores to prevent people from downloading the necessary software. For example, after considerable popularity in 2021 and 2022, Kuaishou, a short video and e-commerce platform, is now unavailable in many MENA region app stores (see Figure 5, below). Users outside of China will not find this version in their app stores.

Type #2. Phone number registration:

Overseas users are not offered the phone number of their country of residence (or the prefix of their SIM card) in the app’s registration interface. When users are unable to select the phone number prefix of their country of residence, receiving a verification message is impossible — turning phone-number-based RNR into an effective access barrier. For example, on “Weibo Intl.,” only residents of 29 countries can sign up (see Figure 5, below).

Type #3. ID verification:

Submitting identification documents and the accompanying loss of anonymity may deter users from registering on a platform if such requirements are believed to be not necessary or appropriate from the perspective of the user.

Type #4. Cross-Account verification:

When account verification can only be completed by verification of other users on a given platform, new user sign-ups may be effectively excluded if they have no prior contacts or choose not to reveal existing ones.

Figure 5: Blocked by numbers – Regional restrictions block Chinese app use

App store level censorship for Kuaishou for users in Algeria, Egypt, Saudi Arabia and others

Phone number registration options on app “Weibo Intl.”

Two different intentions can be identified from these four types of barriers. Types #1 and #2 are exclusionar yby design, representing a hard barrier around the grey zone (see Figure 5). The limiting intent here is obvious and although workarounds exist for these barriers, they do not provide sustainable or reliable solutions.²¹ Types #3 and #4, on the other hand, work to enforce identifiability but do not de facto exclude certain types of users. Instead, these barriers ensure that interactions from new users within the grey zone can be closely monitored. While these barriers pose privacy risks, they are not insurmountable and therefore can be considered soft barriers, in so far as the loss of privacy might discourage new users. The intent here appears to be to permit new users from abroad, so long as they are willing to comply with RNR policies.

Our anonymous survey of at-risk communities indicated that overseas users have indeed encountered each type of access barrier identified above (Figure 6, below). Nearly half of all respondents reported having encountered type #3 and #4 access barriers related to identifiability (49% and 46%, respectively). More than a third reported experiencing barriers of types #1 and #2 aimed at excluding users (38% and 43%, respectively). Survey respondents additionally leaned towards the perception that access barriers on Chinese social media apps were more prevalent as compared to other international competitors. The largest group (45%) reported greater difficulty registering on apps developed by Chinese companies. In contrast, most respondents found account maintenance to be similar on both Chinese and international apps (51%).

Global patterns of access barriers

In scoping the global implementation of transnational hard access barriers of types #1 and #2 (app store-level barriers and unavailable country phone numbers), we selected the 62 most-downloaded Chinese social media apps across 58 countries.²² Our findings reveal that 75% of these apps have implemented phone-number based RNR (as opposed to e-mail registration or free use). This indicates that roughly three-out-of-every-four Chinese-developed apps enforce identifiability requirements for overseas users, representing a significant — and controllable — portion of the grey zone internet (see Appendix 4). In other words, access conditions set by China’s cyber regulators have become almost inescapable.

Further tests were then conducted to determine whether RNR policies have turned into access barriers for information exchange. In doing so, we dropped six of the 39 apps, as they allowed for no exchange of information. The remaining 33 apps represent perhaps the most important information channels in the grey zone internet. We checked hard access barriers for those 33 apps from across 58 countries, resulting in 1914 country-app pairs. At least one access barrier of type #1 or #2 was recorded in 664 instances, representing 35% of possible country-app pairs (see the full overview Access Barrier Matrix in Table 1, below, for more details).²³

Notably, however, the recorded types of hard access barriers and their implementation across the tested apps and regions displayed a considerable degree of heterogeneity — suggesting China’s ideological security firewall is not meant to completely prevent information exchange (or is at least unsuccessful in doing so in practice). Three levels of varying global restrictiveness can be categorized among the tested apps. In the first group of 10 apps (Group 1), the apps are largely unavailable across the world due to either type #1 or type #2 barriers. These apps — which included popular options such as DingDing, Kuaishou, and Douban — either cannot be downloaded at all or users cannot register for them with their national phone numbers (or both).The next group of 15 apps (Group 2) is generally accessible around the world, but occasionally faces both types of hard access barriers. This group also contains some popular apps, including QQ, Weibo (微博), and Zhihu (知乎). The eight apps in Group 3 were freely available across the 58 countries in our tests, and include the likes of WeChat (微信) and Little Red Book (小红书).

The matrix below also provides evidence of geographic distinctions. Certain clusters of apps in Group 2 are only unavailable in particular regions. For example, a set of apps in Group 2 (including QQ and other social networks) are only unavailable in the extended European region, whereas certain video streaming and forum apps (including Kuaishou and Zhihu) are largely inaccessible in the MENA region and South American countries. This pattern is particularly noteworthy in the context of the CCP’s fear of domestic instability and a U.S.-promoted Color Revolution in China.²⁴ The Asia-Pacific region, Sub-Sahara Africa, and countries in Central Asia on the other hand, enjoy fairly unrestricted access, even among apps of Group 1. Perhaps surprisingly, few access barriers exist for apps in the U.S.

Table 1: The Minesweeper of censorship – China limits information exchange on apps

Our survey further underscores the access difficulties experienced across China’s most popular social apps (Figure 7, below). Across 10 major apps, 5% of respondents on average reported being able to download an app but ultimately unable to sign up for it. In addition, survey respondents were unable to download apps roughly 2% of the time, on average. Maintaining access to a registered account also seemed to be an issue, with 7% of respondents reporting they had lost access to their accounts. This issue was particularly prevalent with Weibo (17%), QQ (13%), and WeChat (12%).

Finally, our research also examined how users typically respond to encountering access barriers (Figure 7, below). The majority of respondents adapted to access barriers, with 29% acquiring separate phones dedicated to running only Chinese social media apps, 23% complying with barrier requirements and 8% reporting they did not ultimately mind the barriers when encountered. Nearly one-fifth of all respondents, however, deleted apps affected by access barriers — and were thus effectively deterred from entering the grey zone and participating in China’s online discourse.