Professionals’ Perspectives on AI-Generated Child Sexual Abuse Materials

Background and Purpose. The proliferation of child sexual abuse material (CSAM) has accelerated alongside digital technologies, with recent advances in generative artificial intelligence (AI) introducing new forms of synthetic and manipulated sexual imagery involving children. These developments are not only increasing the scale of CSAM but also transforming its production, distribution, and the nature of victimization. While existing scholarship has focused on technical and legal dimensions, far less is known about how front-line professionals interpret and respond to AI-generated CSAM. This study examines how professionals understand AI-generated CSAM, with attention to shifting conceptualizations of harm, victimhood, and the challenges associated with investigation and prosecution.

Methods. This exploratory qualitative study draws on semi-structured interviews with 21 professionals across 11 U.S. states, including law enforcement personnel, prosecutors, and victim service providers engaged in CSAM-related work. Participants were purposively sampled based on their experience with technology-facilitated CSAM cases. Interviews averaged 50 minutes and were conducted virtually. Using a grounded theory–informed approach, transcripts were coded iteratively. Initial holistic coding of AI-related content was followed by focused thematic analysis to identify patterns in how AI intersects with investigative practices, victim identification, and service provider responses. Collaborative analytic procedures were used to refine themes and enhance rigor.

Results. Findings indicate that professionals are navigating AI-generated CSAM as an emergent and evolving phenomenon characterized by uncertainty and institutional strain. Three interrelated themes emerged. First, participants described shifts in conceptualizations of victimization, emphasizing that generative AI enables identity-based exploitation in which children’s likenesses are appropriated from benign images and transformed into sexualized content without prior contact or grooming. This expands traditional models of abuse and introduces new pathways to harm. Second, ambiguity surrounding victimhood was a central concern. Although participants widely acknowledged the emotional and reputational harm associated with AI-generated CSAM, institutional practices continue to prioritize distinctions between “real” and “synthetic” content, often shaping case triage and service access. This binary was described as inadequate, particularly as offenders “launder” real images through AI to obscure their origins. Third, participants identified significant operational challenges in detection, investigation, and prosecution, including difficulty distinguishing authentic from AI-generated material, limitations of existing forensic tools, and the emergence of “AI defenses” in court that contest the authenticity of evidence. Legal responses were described as uneven and largely untested, contributing to uncertainty in prosecutorial decision-making.

Conclusions and Implications. AI-generated CSAM represents a fundamental shift in the landscape of child sexual exploitation and abuse, challenging established distinctions between real and virtual harm and straining institutional responses. Findings underscore the need for harm-centered frameworks that move beyond binary definitions of victimization to account for identity misuse and psychological impact. Policy and practice implications include the need for clearer statutory guidance, enhanced cross-jurisdictional coordination, and training that integrates technical, ethical, and victim-centered considerations. Additionally, prevention efforts targeting youth and technology developers are critical to addressing emerging forms of misuse. Future research should center victim experiences and evaluate evolving legal and technological interventions to align institutional responses with the realities of AI-mediated harm.