The article by Mitali Thakor explores the intricate sociotechnical ecosystem that enables and shapes global initiatives to detect and regulate child sexual abuse images (CSAI). Rather than framing CSAI solely as prohibited content, Thakor conceptualizes these images as “violent digital objects” – artefacts that circulate through networks of detection, classification, and evidence-building, carrying both the original violence they depict and new forms of violence associated with their handling. Thakor explores how CSAI policing is structured through a wide-ranging investigative infrastructure, composed of law enforcement agencies, private tech companies, international non-profits, content moderators, forensic analysts, and machine-learning systems. These actors collaborate through shared databases, image-hash repositories, and automated detection tools. Yet, Thakor emphasizes, nonetheless, that this infrastructure is not purely technical. It is sustained by extensive human labour, interpretive judgment, and emotional resilience.

A key concept in the article is the role of “algorithmic detectives,” the hybrid assemblages of human reviewers and automated systems used to identify abusive materials. Despite the prominence of tools like PhotoDNA, Thakor argues that automated detection cannot replace human perception. Moderators and analysts remain responsible for training models, deciding ambiguous cases, and contextualizing images – tasks that often expose them to sustained trauma. Thakor also foregrounds racial and political dynamics embedded in CSAI detection. She argues that algorithmic tools and forensic frameworks often rely on datasets that assume whiteness as the normative marker of childhood innocence. This bias shapes what kinds of images are easily recognized, how victims are represented, and which cases receive investigative attention. In this way, CSAI policing can inadvertently reproduce racial hierarchies. Finally, Thakor critiques the expanding reach of surveillance justified through child protection. The emotional and moral gravity attached to child innocence enables states and corporations to extend data collection, monitoring, and information-sharing practices under relatively little scrutiny. While these efforts aim to safeguard children, they also reinforce broader carceral logics and raise ethical questions about privacy, data retention, and cross-border policing.

Thakor’s article is compelling for its combination of ethnographic detail and critical theoretical insight. By framing CSAI as “violent digital objects,” she extends the conversation beyond discussions of content moderation into an analysis of how violence becomes encoded in data infrastructures themselves. Her focus on racialised constructions of innocence is particularly valuable, revealing how ostensibly neutral detection systems can reflect and perpetuate inequality. A potential limitation is that while Thakor critiques the expansion of surveillance and the racial bias of forensic tools, she offers fewer practical pathways toward a correction of this system. Readers may be left wondering how to balance the urgent need for CSAI detection with concerns about privacy, labour exploitation, and over-policing. Additionally, her evidence is primarily drawn from specific Anglo-American contexts, raising questions about how the dynamics she identifies play out across different technological and legal regimes. Overall, the article is an incisive and important examination of the ethical and political stakes of digital child-protection systems.