Special Series: “New Directions in Africa-China Studies” | Yidi Zheng, Capturing the Face: Afro-Inclusive Machine Visions from South China
Introduction
In the latter half of the 2010s, two shifts occurred in the flow of Chinese investment to Africa: one in the object of investment, and the other in the agents who propelled these ventures (Acker and Brautigam 2021; Huang and Pollio 2023). Driven by China’s internal reorientation toward ICT development and African governments’ growing desire to diversify their digital partnerships, the country adeptly repositioned itself—from builder of roads and railways to architect of fiber-optic cables, data centers, and smart cities (Huang and Pollio, 2023; Soulé, 2025). Simultaneously, Chinese commercial banks and private firms notably expanded their roles in financing digital connectivity projects. In contrast, loan commitments from traditional policy banks, such as China Exim, began to show signs of deceleration (Acker and Brautigam 2021). As elucidated by Huang and Pollio (2023), this evolution signifies a broader pivot away from the state-backed infrastructure paradigm toward new configurations of “market-in-state” logics that now underpin the digital turn.
What emerged alongside the mixed-funding digital infrastructure, slowly yet steadily across the African continent, were the facial recognition technologies (FRTs) that were only made possible by the digital connectivity afforded by such infrastructure (Pollio 2024). This essay revisits the pivotal year of 2018, when two southern Chinese tech companies, Transsion Holdings (hereinafter referred to as Transsion) and CloudWalk Technology (hereinafter referred to as CloudWalk), rolled out their respective FRTs to African markets and governments. Though distinct in their approaches to integrating FRTs into their product ecosystems and relationships with state funding, both entries into Africa were seen as historic by cautious scholars and anxious market observers alike (Avle 2022; Travers 2024; Feldstein 2019).
When Transsion launched its flagship AI-enhanced camera, which was pre-installed on one of its subsidiary brands, Tecno, it had already captured over 30% of the African smartphone market, doubling that of the second-place holder, Samsung (Counterpoint 2019). Dubbed the “smartphone king of Africa” (Nyabiage 2024), Transsion’s rise is often narrated as a story of hard-earned market acceptance that helped to reshape perceptions of Chinese innovation across the continent (Pollio 2024). CloudWalk’s presence, by contrast, appears contentious and unappealing. Its 2018 entry into Zimbabwe via a state-negotiated AI surveillance deal provoked public concern and international scrutiny (Travers 2024). One firm was driven by market enthusiasm; the other by a president keen to curtail political dissent.
What binds them, however, is more than just origin or timing. It is their shared commitment to better recognize blackness and develop FRTs that can see, detect, and differentiate darker-skinned faces (Avle 2022; Mudzingwa 2018). The racial bias of machine vision systems, long exposed for privileging lighter skin due to unbalanced training datasets, has generated urgent calls for algorithmic justice. For some, this means diversifying the datasets and including a broader range of phenotypical representations (Buolamwini and Gebru 2018). For others, the demand for inclusion is itself part of the problem (Benjamin 2019; Amaro 2016).
In his e-flux essay “As If” (2016), Ramon Amaro identifies two central problems with integrating visual data of Black life into algorithmic systems. First, the machine must produce a “fictive and compulsive ordering of human attributes into a single coherent image of species,” which makes Black Africans computationally identifiable. Second, this coherence risks disconnecting the algorithmic image from its historical categories, stereotypes, and the lived realities of Black lives (Amaro 2016). While Amaro’s (2023) task is to imagine the possibilities of a human-techno relation that goes beyond the desire for recognition, such desire, driven by tech companies, nation-states, and the citizen-users of digital infrastructure and devices, is steadfastly realizing the inclusion of blackness into the machine vision of FRTs.
Re-imagining a techno-human relation beyond the bounds of current racial and algorithmic logic requires understanding what kinds of techno-human relations are embedded in the design of the most widespread machines. Using the cases of Transsion and CloudWalk, I investigate the companies’ technological and rhetorical claims to inclusive recognition and reveal the hidden forms of norms and power made invisible in the process of designing Afro-Inclusive FRTs. Building on Amaro’s critique of racial coherence in machine vision, I contend that the project of inclusive facial recognition turns Black African subjects into training data, whose facial attributes are segmented, stabilized, and re-rendered to conform to the logic of machine vision and algorithmic legibility. Their corporate claims to inclusivity also serve to detach Chinese facial recognition technologies from their own epistemic conditions while obscuring the abuses of power and racialized relations they enact on the ground.
Literature and Method
Across disciplines, much of the literature on China’s infrastructural investment in African countries, whether physical or digital, approaches the topic through the lenses of colonialism, coloniality, or frontier capitalism. Even as the digital sector’s actors have grown more diverse and fragmented compared to traditional infrastructure, discussions of China’s Belt and Road Initiative (BRI) and Digital Silk Road (DSR) continue to frame loans and bilateral deals as potential neo-colonial forces. Private companies exporting FRTs, as Ishan Sharma (2020) suggests, must be understood in concert with state institutions, forming what he terms a security-industrial complex. As the world’s largest exporter of AI surveillance systems, with FRTs as a subset, China has undeniably leveraged credit lines and loan-linked contracts to expand the reach of firms like Huawei, ZTE, and Hikvision across the African continent (Feldstein 2019).
However, continent-to-state scale analyses risk obscuring the specific ways in which individual African countries and their internal dynamics shape digital infrastructure deals (Huang and Pollio 2023; Soulé 2025; Otele 2020). A blanket neo-colonial diagnosis, for instance, would flatten the distinction between African-initiated contracts (as in Kenya) and debt-enforced procurement (as in Mauritius) (Feldstein, 2019; Huang & Pollio, 2023). Conversely, scholars focusing on Global China and South–South relations often add nuance, presenting companies like Transsion as alternatives to the white-coded design logics of Euro-American technologies (Avle 2022; Lu 2022). This essay, however, joins Sareeta Amrute’s (2025) critique of applying the framework of “the Global South” in discussions of AI. She argues this term often obscures the relations of exploitation within reconstituted national territories. Instead, she proposes the adoption of “the majority world,” an analytic that inherently recognizes the “power differentials within the so-called ‘developing world,’ and forms of oppression that keep non-elite ways of knowing and living from entering into our decision-making about collective presents and futures” (Amrute 2025).
In this spirit, my inquiry does not aim to label Chinese FRTs in Africa as colonial exports or to celebrate them as liberatory South–South technology. Rather, I focus on the forms of racial knowledge that are generated, optimized, and circulated in the production of Afro-inclusive FRTs. My method of desk research is informed by Acker and Brautigam’s (2021) appropriation of the term “forensic internet sleuthing.” I examine a corp of publicly available materials produced or commissioned by CloudWalk and Transsion, including research papers and internal documents. I also analyse corporate rhetoric through product launches, visual demonstrations, user product reviews, and public statements by CloudWalk and Transsion engineers and executives.
Data Capture
Both born in the southern part of China, Transsion and CloudWalk garnered vastly different reputations for their introduction of facial recognition technologies (FRTs) to African countries. In April 2018, Transsion released its camera-centric smartphone models Camon X and Camon X Pro. Prior to this release, Transsion’s feature phones and smartphones had already secured over 30% of the phone market in African countries with their “user-centered design,” outperforming long-time competitors like Nokia, Samsung, and Motorola (Pollio 2024). The addition of facial recognition and AI camera systems optimized for darker-skinned users became a key appeal of Transsion’s products across the continent (Lu and Qiu 2022). In contrast, CloudWalk’s export of its AI facial recognition technology to Zimbabwe was met with widespread public grievance and critique (Travers 2024).
Cloudwalk and Transsion’s FRTs differ also in the hardware in which their algorithms are embedded. Cloudwalk’s self-developed three-dimensional structured lighting face recognition technology was designed to reside across Cloudwalk’s own assemblage of Infrared Binocular Camera, Facial Recognition Access Control System, IBIS Integrated Biometric Identification System, Identify Verification Engine, and AI Big Data Analytics and Platform (Gu 2018). Transsion’s FRT, on the other hand, is built into its smartphone cameras, aiding fast Face ID unlock, AI-powered selfie beautification, and automatic white balance, among other features (Mátùlúkò 2018)—all optimized for photographing dark-skinned users.
Beyond different material embeddedness, the two companies source their training data required to adopt FRTs for richly-melonated users via vastly different channels. While Cloudwalk initially trained its export FRT with self-collected data, its 2018 agreement with the Zimbabwean government led to the mass collection of profiling data of Zimbabweans, categorized by lighting conditions in Zimbabwe and phenotypical variation of its population (Zhang 2018; Haoqixin Ribao 2018). Although it has been reported to have analyzed “several million photos of dark-skinned Africans” (Chutal 2022), Transsion’s data collection methods remain unclear. Based on ethnographic accounts and research reports, it is safe to assume that Cloudwalk collects training data through both internal collection and research initiatives. However, the extent of both activities cannot account for the amount of training data accessed. One Transsion employee reported that photos of mixed-race Transsion employees were taken and submitted to headquarters because internal photo acquisition does not raise consent issues (Pollio 2024). Transsion’s recent research collaboration with the University of Leeds, in contrast, represents a consent-conscious approach for the company to collect skin reflectance data from people of various skin tones (Zhou et al, 2025).
After Capture
Despite these contrasting receptions, Transsion and CloudWalk are connected by their shared investment in inclusive machine vision. Nevertheless, the “inclusion” they offer does not merely insert blackness into a historically exclusive regime. I argue that it transforms Black Africans into training data, segmenting, stabilizing, and re-rendering their facial attributes according to the logics of hardware calibration and algorithmic legibility.
Unlike Transsion’s sustained self-positioning to be the vanguard of inclusive algorithms for historically excluded Black faces, CloudWalk appeared to have quietly walked back from its promise to adopt its FRTs for African facial morphologies, at least in the English-speaking world. While academic journals and foreign news outlets continued to report on the data transfer between the Zimbabwe government (see Chutel 2022; Okolo et al. 2023), the key source confirming this data transfer is a now-deleted Global Times interview with CloudWalk R&D director Yao Zhiqiang that could only be retrieved via the Internet Wayback Machine. What was also shrouded in the disappearance of this interview was CloudWalk’s language on training African faces with the data provided by the Zimbabwe government. This eschewance is likely a result of the anxieties and warnings over data privacy from within Zimbabwe and out of the United States, whose Department of the Treasury enlisted CloudWalk as an extension of China’s military industrial complex subject to sanction, citing its data transfer deal with Zimbabwe alongside its activities in Xinjiang and Tibet (2021). Although the language of racial differences and data transfer remains in Chinese journalistic reports (Haoqixin Ribao 2018), CloudWalk’s official language now refocuses on their three-dimensional structured lighting FRT’s ability to accurately identify masked individuals (CloudWalk 2022).
Publicly available information on CloudWalk and the Zimbabwean government’s collaboration reveals the general process of data collection and circulation: 1) collection and digitization of citizens’ facial data and personal information by the Zimbabwe government; adaptation of CloudWalk FRT algorithm for dark-skinned individuals and integration of FRT algorithm and national data platform; 3) deployment of CloudWalk hardware cameras and sensors in Zimbabwe’s border checkpoints, urban surveillance systems, and police databases (Haoqixin Ribao 2018; Gu 2018; Travers 2024). As outlined by a Quartz (2022) reporter, CloudWalk’s advanced three-dimensional structured lighting technology is uniquely equipped to recognize dark skin tones. Since 3D structured light captures the depth and contours of facial data, rather than color and pigmentation, it likely improves recognition results when trained on images of dark skin; however, apart from the initial disclosure of data acquisition from the Zimbabwe government, CloudWalk remained reticent about the afterlives of the transferred data sets, as well as the training process of CloudWalk’s inclusive FRT.
Through a reading of Transsion’s research output, we see the way face and skin tone are categorized, calibrated, and stabilized to achieve realistic skin color across all skin tones. This aforementioned research article, co-authored by Transsion engineers and their collaborators from the University of Leeds, explores the development of the SCR-AWB (Spectral Color Recovery Auto White Balance) algorithm. This algorithm, when installed on a smartphone camera, detects and identifies skin-dominant areas before adjusting the white balance for optimal image reproduction (Zhou et al. 2025). Facial regions are then weighted by contrast and texture sensitivity, with metrics such as gradient magnitude, Laplacian values, and local phase coherence used to emphasize visually areas, including the eyes, nose, and cheeks (Zhou et al. 2025). The researchers use learned weights from inclusive datasets that include darker skin tones. In doing so, the face becomes a site of perceptual calibration, where selected facial zones are rendered into sites of algorithmic intervention to meet machine vision’s standards of visual salience. As algorithmic systems enforce a false coherence that regulates visibility, Amaro notes, they interpellate individuals “only inasmuch as [they] can be measured against a universalizing concept of being” (2016).
Faces of Inclusion
The rhetoric of inclusivity permeates both Transsion’s and CloudWalk’s public-facing language. From product launch statements to executive interviews, the claim that their technologies are made “for Africa” recurs with disciplined regularity. Tecno, a Transsion subsidiary, claims that it designs phones “in line with Africa’s needs” (African Business, 2020), while its engineers describe the development of a “multi-skin tone imaging system” as a gesture of ethical innovation (PR Newswire 2023). The claims of inclusivity and of correcting the wrongs of Western algorithmic design, I argue, serve both to detach neutralized Chinese technology from its own normative assumptions and to obscure the exercise of power and racial relations present on the ground.
The first detachment is the severing of inclusive claims from the technical realities embedded in design. The optimization of Black faces in FRTs occurs not through an open embrace of diversity, but through a regime of selection and correction. Skin brightness, edge texture, and visual sharpening are calculated and adjusted until the face conforms to a model recognizable to the system and to its engineers. Tecno’s facial enhancement algorithms do not just accurately capture African skin tones. YouTube reviewers of Tecno’s AI beautification feature repeatedly show how the camera slightly lightens the skin (Fred’s Tech Hub 2020; Valor Reviews 2020). These effects, which appear seamless to the user, are calibrated defaults rooted in an aesthetic standard aligned with beauty filters that have been normalized from China to Africa. This system of enhancement does not emerge organically from user feedback, but stems from the normative assumptions of Chinese engineers and product managers at FRTs. Anti-racist technoscience, as Ron Eglash states, can sometimes be asserted “by excluding one type of value as a disguise for others” (Eglash 2019).
The second detachment falls between rhetorical inclusion and lived exclusion. Despite the official rhetoric of bringing security and modernity to Zimbabwe, the real-life deployment of facial recognition surveillance cameras neither targeted areas with high crime rates nor resulted in the conviction of any criminals (Travers 2024). On the contrary, local informants identified “opposition party members, civil society leaders, and citizens who use social media for political expression” as groups targeted by CloudWalk-powered surveillance cameras (Travers, 2024), thereby excluding them from political and civic participation. CloudWalk’s racial optimization model was developed in Guangzhou, a city that has hosted hundreds of thousands of African migrants for over two decades. However, during China’s Zero-COVID years, these same African residents faced what Castillo and Amoah (2020) describe as a “pandemic-era regime of biometric exclusion.” Africans, in particular visa overstayers, were locked out of apartments, denied access to shops, and rendered ineligible for the Health Code system that governed mobility across the city. They became invisible to the apps, but hyper-visible to the state.
Amaro (2016) critiques algorithmic inclusion for how it conceals “the algorithm’s reliance on historical category, namely, what features represent the categories of human, gender, race, sexuality, and so on.” However, here, the obfuscation is not only historical. Modern forms of normative assumptions, even if not explicitly racial, still play a significant role in the design and operation of FRTs; contemporary relations of discrimination and hierarchy end up hidden behind the progressive efforts of inclusive machine vision.
Conclusion
Once characterized by highways, railways, and mining zones, the landmarks of China-Africa engagement have gradually drifted into the digital infrastructure, encompassing fiber optic cables, AI systems, and biometric platforms. The agreements forged between CloudWalk and the Zimbabwean government not only represent a move within China’s Belt and Road Initiative but are also aligned with Zimbabwe’s Smart Zimbabwe 2030 Master Plan (Travers 2024). Yet, amid the nation-state actors, this digital transformation has come with a shift in agents from policy banks and state-owned enterprises to private firms like CloudWalk and Transsion, whose overseas strategies are animated not just by national strategy but also by technical ambitions and commercial experimentation.
In this new frontier, articulations of inclusive algorithms become a 21st-century addition to the long line of PRC rhetoric of mutual benefit and non-interference in China-Africa relations (Strauss 2009). Like many of its predecessors, this essay demonstrates that the rhetoric of inclusive algorithms does more than it claims. Racial inclusivity in machine vision requires not only the segmentation, calibration, and reification of Black facial features but also further detaches the technologies from the political and racial relations that structure them. This preliminary report aims to serve as a call for critical scrutiny of present and future upcoming technological intermediations and the attached rhetorical narratives within China-Africa’s digital transformation.
Bibliography
Acker, Kevin, and Deborah Brautigam. 2021. “Twenty Years of Data on China’s Africa Lending.” Washington, DC: SAIS China-Africa Research Initiative.
Amaro, Ramon. 2016. “As If.” E-Flux. 2016. https://www.e-flux.com/architecture/becoming-digital/248073/as-if.
Amaro, Ramon. 2023. The Black Technical Object: On Machine Learning and the Aspirations of Black Being. Cambridge, Massachusetts: MIT Press.
Amrute, Sareeta. 2025. “Thinking the Unthinkable in AI: Four Hegemonic Ways of Seeing and Five Majority World Ways to Move beyond Them.” Antipode, 1–23. https://doi.org/10.1111/anti.70051.
Avle, Seyram. 2022. “Hardware and Data in the Platform Era: Chinese Smartphones in Africa.” Media, Culture & Society 44 (8): 1473–89. https://doi.org/10.1177/01634437221128935.
BellaNaija. 2018. “TECNO Releases the Camon X and X pro with 24MP Cameras & New Face ID Technology + F Series in Partnership with Google.” BellaNaija. 2018. https://www.bellanaija.com/2018/04/tecno-releases-camon-x-pro-f2-series.
Benjamin, Ruha. 2019. “Introduction: Discriminatory Design, Liberating Imagination.” In Captivating Technology: Race, Carceral Technoscience, and Liberatory Imagination in Everyday Life, edited by Ruha Benjamin, 1–24. Durham: Duke University Press.
Buolamwini, Joy, and Timnit Gebru. 2018. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification.” In Proceedings of the 1st Conference on Fairness, Accountability and Transparency. Vol. 81. 77-91: PMLR.
Burt, Chris. 2018. “Implementation of CloudWalk Facial Recognition Technology in Zimbabwe Progressing in Stages.” Biometric Update. 2018. https://www.biometricupdate.com/201805/implementation-of-cloudwalk-facial-recognition-technology-in-zimbabwe-progressing-in-stages.
Cahill, Susan. 2023. “Surveillance Frontierism: Art and the Colonial Project of Surveillance.” Media Practice and Education 24 (2): 128–46. https://doi.org/10.1080/25741136.2023.2212099.
Castillo, Roberto, and Padmore Adusei Amoah. 2020. “Africans in Post-COVID-19 Pandemic China: Is There a Future for China’s ‘New Minority’?” Asian Ethnicity 21 (4): 1–6. https://doi.org/10.1080/14631369.2020.1773760.
Chutel, Lynsey. 2022. “China Is Exporting Facial Recognition Software to Africa, Expanding Its Vast Database.” Quartz. 2022. https://qz.com/africa/1287675/china-is-exporting-facial-recognition-to-africa-ensuring-ai-dominance-through-diversity?.
CloudWalk 云从科技. 2024. Yuncong keji jituan gufen youxian gongsi 2024 nian niandu baogao 云从科技集团股份有限公司2024年年度报告 [2024 Annual Report of CloudWalk Technology Co., Ltd.]. CloudWalk.
Deck, Andrew. 2020. “Your Guide to Transsion, Africa’s Biggest Mobile Phone Supplier.” Rest of World. 2020. https://restofworld.org/2020/transsion-from-china-to-africa.
Dixon-Román, Ezekiel, and Ramon Amaro. 2021. “Haunting, Blackness, and Algorithmic Thought.” E-Flux Journal, no. 123.
Eglash, Ron. 2019. “Anti-Racist Technoscience: A Generative Tradition.” In Captivating Technology: Race, Carceral Technoscience, and Liberatory Imagination in Everyday Life, edited by Ruha Benjamin, 277–251. Durham: Duke University Press.
Feldstein, Steven. 2019. “The Global Expansion of AI Surveillance.” Carnegie Endowment for International Peace.
Gu, Wenli 顾文立. 2018. “Xiang Feizhou chukou heikeji Zhongguo ‘yingyan’ jiang fuwu Jinbabuwei” 向非洲出口黑科技 中国“鹰眼”将服务津巴布韦 [Exporting Black Tech to Africa: China’s ‘Eagle Eye’ Will Serve Zimbabwe]. ImSilkroad. https://www.imsilkroad.com/news/p/92122.html.
Haoqixin Ribao 好奇心日报. 2018. “Zhongguo zhengzai xiang Feizhou shuchu renlian shibie jishu, 目前处于建数据库阶段”中国正在向非洲输出人脸识别技术，目前处于建数据库阶段 [China Is Exporting Facial Recognition Technology to Africa, Currently in the Database-Building Stage]. D1net. https://www.d1net.com/bigdata/database/528647.html.
Huang, Zhengli, and Andrea Pollio. 2023. “Between Highways and Fintech Platforms: Global China and Africa’s Infrastructure State.” Geoforum 147. https://doi.org/10.1016/j.geoforum.2023.103876.
Lu, Miao. 2021. “Translating a Chinese Approach? Rural Distribution and Marketing in Ghana’s Phone Industry.” Media, Culture & Society 43 (2): 309–25. https://doi.org/10.1177/0163443720987811.
Lu, Miao. 2022. “Technology Translations between China and Ghana: The Case of Low-End Phone Design.” In Critiquing Communication Innovation: New Media in a Multipolar World, edited by Rolien Hoyng and Gladys Pak Lei Chong. East Lansing, Michigan: Michigan State University Press.
Lu, Miao, and Jack Linchuan Qiu. 2022. “Empowerment or Warfare? Dark Skin, AI Camera, and Transsion’s Patent Narratives.” Information, Communication & Society 25 (6): 768–84. https://doi.org/10.1080/1369118x.2022.2056500.
Mátùlúkò, Múyìwá. 2018. “Everything We Know about the TECNO Camon X and 4 Other Phones Announced Yesterday.” Techpoint Africa. 2018. https://techpoint.africa/general/tecno-camon-x-features.
Mudzingwa, Farai. 2018. “Mnangagwa’s Govt Getting Facial Recognition Tech from China.” Techzim. 2018. https://www.techzim.co.zw/2018/04/mnangagwas-govt-getting-facial-recognition-tech-from-china.
Nyabiage, Jevans. 2024. “Alarm Bells Ring for ‘Smartphone King of Africa’ as Other Chinese Brands Chase Crown.” South China Morning Post, 2024. https://www.scmp.com/news/china/diplomacy/article/3289680/alarm-bells-ring-smartphone-king-africa-other-chinese-brands-chase-crown.
Okolo, Chinasa T., Kehinde Aruleba, and George Obaido. 2023. “Responsible AI in Africa—Challenges and Opportunities.” In Responsible AI in Africa: Challenges and Opportunities, edited by Damian Okaibedi Eke, Kutoma Wakunuma, and Simisola Akintoye, 35–64. Switzerland: Springer Nature.
Otele, Oscar Meywa. 2020. “China’s Belt and Road Initiative and Intra-Regional Dynamics in Africa.” African Studies Quarterly 19 (3-4).
Pollio, Andrea. 2024. “Of Bloatware and Spreadsheets: Nairobi, Chinese Phones, and the Limits of Data Coloniality.” Journal of Urban Technology, August, 1–22. https://doi.org/10.1080/10630732.2024.2374195.
PR Newswire. 2023. “TECNO Unveils ‘Universal Tone’ - a Multi-Skin Tone Imaging Technology Reflecting Everyone in One Picture.” PR Newswire. 2023. https://www.prnewswire.com/apac/news-releases/tecno-unveils-universal-tone--a-multi-skin-tone-imaging-technology-reflecting-everyone-in-one-picture-301971264.html.
Sharma, Ishan. 2020. “China’s Neocolonialism in the Political Economy of A.I. Surveillance.” The Cornell International Affairs Review XIII (Spring): 4–64.
Soulé, Folashadé. 2025. “Negotiating Africa’s Digital Partnerships amid Geopolitical Competition.” Centre for International Governance Innovation. www.cigionline.org/publications/negotiating-africas-digital-partnerships-amid-geopolitical-competition.
Strauss, Julia C. 2009. “The Past in the Present: Historical and Rhetorical Lineages in China’s Relations with Africa.” The China Quarterly, no. 199: 777–95. https://www.jstor.org/stable/27756501.
Tang, Flora. 2019. “Transsion Holdings’ Journey to an IPO: Strengths and Challenges.” Counterpoint Research. https://www.counterpointresearch.com/insight/transsion-holdings-journey-ipo-strengths-challenges?.
Travers, L. 2024. “Algorithmic Coloniality? The Case of Chinese Artificial Intelligence Technology and Zimbabwean Surveillance.” Transcience 15 (2): 48–82.
U.S. Department of Treasury. 2021. “Treasury Identifies Eight Chinese Tech Firms as Part of the Chinese Military-Industrial Complex.” U.S. Department of the Treasury. December 16, 2021. https://home.treasury.gov/news/press-releases/jy0538?.
Yedder, Omar Ben. 2020. “TECNO Provides Innovations in Line with Africa’s Needs.” African Business. 2020. https://african.business/2020/11/technology-information/tecnos-success-is-built-on-providing-innovations-in-line-with-africas-needs.
Zhang, Hongpei. 2018. “Chinese Facial ID Tech to Land in Africa.” Global Times. 2018. https://web.archive.org/web/20180607044026/https://www.globaltimes.cn/content/1102797.shtml.
Yidi Zheng is a PhD candidate in Cultural Anthropology at Duke University. Her research concerns the impact of digitized urban development on African migrant communities in southern China.
To cite this essay. please use the bibliographic entry suggested below:
Yidi Zheng, “Capturing the Face: Afro-Inclusive Machine Visions from South China,” criticalasianstudies.org Commentary Board, January 22, 2026; https://doi.org/10.52698/MRTP8026.