Prevalence of technology-facilitated child sexual exploitation and abuse (TF-CSEA)
Measuring TF-CSEA in representative surveys is a nascent but growing evidence base.
We found an increase by 19 studies published between 1 October 2023 and 31 December 2024 using nationally or sub-nationally representative sampling that measure some form of TF-CSEA prevalence. The most frequently measured data is past year data on online solicitation, child sexual abuse material (CSAM)/ imagebased sexual abuse (IBSA) and exposure to unwanted sexual content with good country coverage across regions.
Childhood lifetime estimates (experience before the age of 18) are less frequently measured than past year estimates likely due to the changing nature of technology across time. Online sexual exploitation and online sexual extortion are captured in far fewer representative surveys, suggesting a need to enhance ways and opportunities to measure these types of potential harms. Many included studies are starting to disaggregate by sex but do not include prevalence by perpetrator type.
The most frequently occurring type of TF-CSEA captured in surveys is online solicitation which was self-reported by 26.6% of participants that they had experienced at some point in their childhood.
Online solicitation covers a range of unwanted or pressured sexual interactions, which may include casual sexual inquiries via mobile phone or the internet, or long-lasting sexual conversations that can lead to the exchange of sexual texts/pictures/videos or exposure of intimate body parts. All types of online solicitation may come from peers as well as adults.
Online solicitation is defined broadly and diversely here, given the infancy of the field and the various kinds of questionnaire formulations that have been used. For example, some of the surveys specify that children received unwanted sexualised messages personalised for them, but others include encountering this type of messages that may have been directed at many respondents. It is important to note that majority of the surveys did not report perpetrator type, thus all different types of online solicitation may come from peers as well as adults. The diminished capacity to identify perpetrators in the online environment results in limited data on their characteristics. Incorporating follow-up questions on the child’s subjective perception of the perpetrators would allow more accurate identification and classifications of these unwanted behaviours. Once more consistent approaches to research and data collection are developed and more data are available on the impact of those behaviours, more granular and precise classifications will be possible. However, until then, the assemblage of different questionnaire items referring to this subtype of TF-CSEA identified in this research effort affords a basis for a more formal study of key definitional components.
There is an ongoing debate about how to best define online solicitation as a form of online sexual abuse and its consequences for epidemiology (Bulger et al., 2017; Finkelhor et al. 2024). Online solicitation is also increasingly illegal in national legislation and defined by major child protection bodies such as the Lanzarote Commission, UNICEF, ECPAT and Childlight as a form of TF-CSEA.
In the absence of universal agreement about equating all forms of online sexual solicitation with child sexual abuse, due to variability in intent, context and harm, Childlight adopts a broader definition of TF-CESEA that includes unwanted sexual communication. This reflects emerging international frameworks and consistent evidence that such experiences can be harmful and may form part of pathways to more severe exploitation.
For past year experiences, we see unwanted exposure to sexual content (7.3%) and online solicitation (6.7%) reported by children in surveys.
It is not surprising that these two types of abuse are most frequently reported as they encompass a range of potential harms, with varying protections in place for children depending on the country context.
For the smaller number of studies that disaggregate by sex, we see different types of TF-CSEA being reported by males and females.
In existing representative data, we see that females report higher prevalence for online solicitation, online sexual exploitation and sexual extortion, including experiences both during childhood and in the past year. Whereas, males report more unwanted exposure to sexual content and CSAM/IBSA in the past year and more unwanted exposure to sexual content during their childhood compared to females.
There is variable coverage by UNICEF regions on the number of representative surveys exploring different subtypes. The more frequently measured subtypes include online solicitation, CSAM/IBSA and exposure to unwanted sexual content.
East and Southern Africa, Latin America and the Caribbean and Western Europe are regions that report higher prevalence of online solicitation (past year recall), while Eastern Europe and Central Asia and Western Europe show higher prevalence for past year exposure to unwanted sexual content.
Regional variation in reported online sexual solicitation and unwanted sexual exposure likely reflects a combination of differences in children’s digital access, platform use and risk environments, rather than purely underlying differences in victimisation. For example, higher levels of unwanted sexual exposure in some regions may be associated with earlier and more widespread digital access, as well as platform ecologies characterised by high use of large-scale social media, videosharing, and image-based platforms, which increase the likelihood of encountering sexualised content. These patterns may also reflect higher disclosure rates and greater awareness of online risks, rather than exposure alone. However, there remains much to be understood about regional variation, including how technologyfacilitated child sexual exploitation and abuse is experienced, interpreted and captured in both child and retrospective adult survey data.
South Asia, West and Central Africa and Middle East and North Africa regions are lacking in representative survey data.
This limits data insights on TF-CSEA to CSAM data only with two of these regions showing high CSAM volume and rates (see section Regional Analysis: The What, Where and How of CSAM data). However, more population-level measurement on victimisation experiences during childhood is needed to more fully understand and prevent TF-CSEA.
Scale and nature of child sexual abuse material (CSAM)
Child sexual abuse material data, collected from a few of the major CSAM tracking and analysis organisations globally, can provide insight into both the scale of its availability and the changes in its nature. By bringing the data generated and reported on from each organisation together it can help to provide the global picture of CSAM. Despite many limitations, the data continue to show that there is value in examining the similarities and differences between organisations, without which we would have less knowledge as to the perpetration and offending against children associated with CSAM availability.
CSAM rates, defined as the number of child sexual abuse images and videos per country proportionate to population size, remain high year-on-year in North America and Western Europe.
The child sexual abuse material (CSAM) rate calculated by Childlight combines both hosting and reporting data and standardises these using total country population. Using total population, rather than child population alone accounts for CSAM involving broader systems and actors beyond children themselves. This is also a commonly considered factor in epidemiology, which, until we have stronger evidence of the multiple influences on CSAM availability, has been used to provide global comparability.
Using this approach, high-income regions such as North America and Western Europe show increasing CSAM rates relative to population, compared to declining trends in other regions. These patterns may reflect differences in technological access, detection capacity and reporting practices, but may also indicate a higher relative burden on public sector and regulatory systems.
Our report is one of the first studies to include new global data as well as frontline sector insights into CSEA Guidance Materials and the instructional, justifying and normative influencing nature of this content.
These materials, commonly referred to as ‘paedophile manuals’, though more accurately described as child abuse guidance materials, are poorly defined in law and underexamined in research, despite their recognised role in offender behaviour and their inclusion in legislation in several jurisdictions. Global data specifically from peer-to-peer file sharing networks highlight that this material is circulating across at least 61 countries. In-depth discussion with experts in identifying and addressing this material highlights that these guidance materials are classified by their instructional nature, their justifications for legitimising CSEA and their creation of normative communities that perpetrate CSEA. It is important to note that these guidance materials can be used by children and adults against children, known or unknown; with experts highlighting that possession of this guidance material often coincides with possession of CSAM and/or contact offending against children suggesting this is an important area for policy and prevention consideration by countries.
Commercial CSAM tags show higher than average tags across the content hosted in the Eastern Europe and Central Asia region.
Across 2023 and 2024, Eastern Europe and Central Asia exhibited consistently higher than average levels of tagged commercial CSAM (images/videos which are associated with commercial exploitation or trade), with 6.4% of all identified CSAM from one data source classified as commercial in 2023, decreasing to 2.5% in 2024. Beyond these elevated proportions, the region also recorded the highest absolute volume of commercial abusive content within the data, exceeding other regions by more than 1,000 images per year. Further research could explore this more to understand the underlying drivers of this finding and why it is hosted more often in this region.
There is an increase in the proportion of CSAM images and videos that is being tagged by analysts as ‘self-generated’ content.
‘Self-generated’ CSAM is a type of media showing individuals who have physical control of their recording device (i.e., selfies, self-recordings from their computers, etc.), which may have been shared directly or captured indirectly by other means. This can be created due to the grooming, deception or extortion of a child by an offender which can also be another child or an adult. It may also stem from a consensual correspondence between peers or coercive communication that becomes nonconsensual in its sharing or continued possession. Due to lack of agreement on preferred terminology, we have used single quotes throughout the document to note the limitations of this terminology.
Looking at one data source across several countries, we see an increase by 40-65% of ‘self-generated’ content that was tagged from 2023 to 2024. This highlights that this type of content is increasingly being recognised within content analysis, which is important. It also highlights the need for increasing access to report-and-remove type services such as Take It Down operated by the National Centre for Missing and Exploited Children (NCMEC), Report Remove operated by the Internet Watch Foundation (IWF) among others, which allows children and young people to report content they think is circulating online for removal.
We continue to see content tagged as AI-Generated or Virtual CSAM.
From one new data source, we can see what is echoed in other data sources and that is the continued emergence of the virtual or AI-generated content. This is increasingly being tagged by analysts particularly in North America. As with other tagging data, it may be that those assessing the content are getting better at identifying and documenting the content as opposed to increases in exposure alone. But what we know is that analysts and those at the frontline are increasingly seeing this content which translates to a myriad of challenges and efforts to address this emerging type of content, including curating AI-CSAM specific hash sets, though this may not be sustainable given the high volume of CSAM AI can help produce.