Case study: Sexual recordings of 3-6-year-olds via online devices

The term ‘self-generated’ imagery refers to images and videos created using handheld devices or webcams and then shared online. Children are often groomed or extorted into capturing images or videos of themselves and sharing them by someone who is not physically present in the room with them, for example, on live streams or in chat rooms. Sometimes children are completely unaware they are being recorded and that there is then a image or video of them being shared by abusers.  Many of these images are taken at home in children’s bedrooms or in family bathrooms when the child is alone or with another child such as a sibling or friend.  

'Self-generated' material is something that has risen year on year and a trend we are constantly monitoring. We are now seeing much younger children appearing in this type of abuse imagery. For this reason, we took a closer look to provide an insight into what’s going on.  



  1. We performed an analysis based upon individual image hashes – digital fingerprints – which were tagged as containing a 3–6-year-old child, and 'self-generated'. 

    Please note that an image assessment is an individual assessment based on one image at a time. This is different to when we report on 'report' or 'URL' or 'webpage' assessments, which may contain many images within one report, or URL/webpage. 

  2. We additionally viewed a sample of images which fell within the scope of the study to report what we could see happening in the images.  


Findings based on hashed image analysis

Of the 2,401 'self-generated' images and videos of 3–6-year-olds that we hashed this year, 91% were of girls and most (62%) were assessed as Category C by our analysts. These images showed children in sexual poses, displaying their genitals to the camera.  

The full assessment breakdown is shown in the chart.  

'Self-generated' images of 3-6-year-olds by severity

We assess child sexual abuse material according to the levels detailed in the Sentencing Council’s Sexual Offences Definitive Guideline. The outline of the different categories of child sexual abuse material are:

  • Category A: Images involving penetrative sexual activity; images involving sexual activity with an animal, or sadism. 
  • Category B: Images involving non-penetrative sexual activity. 
  • Category C: Other indecent images not falling within categories A or B. 

Where Category B material was seen, the children were typically rubbing genitals (categorised as masturbation) using their hands/fingers or, less often, another object, such as a pen or hairbrush. 

There were 356 Category A, 'self-generated' images or videos of 3–6-year-olds hashed this year. Most of the Category A material involved children penetrating themselves, or another child.  

The idea that a 3–6-year-old child has unsupervised access to an internet enabled device with camera will be a shock to many people, however, the fact that young children are easily manipulated by predators will be no surprise.  

While children grow up, it is quite normal for there to be an element of sexual experimentation and body-curiosity; that is not what we find in these 'self-generated' images and videos of child sexual abuse. To be clear, the term 'self-generated' does not mean that the child is instigating the creation of this sexual content themselves, instead they are being groomed, coerced and in some cases blackmailed into engaging in sexual behaviour. They are never to blame for the creation or existence of this imagery.  

This is, of course, particularly the case for the age group we are looking closer at in this study. 3-6-year-old children are sexually naive and would not normally be aware of the possibility of this type of sexual behaviour without someone else telling them or showing them what to do. They are easily manipulated and are therefore an easy target for predators who are looking to exploit them.  


Metadata analysis

IntelliGrade, our bespoke grading software, allows us to add additional contextual metadata as we grade and hash the material. These extra tags describe the sexual activity seen and enable our assessments to be compatible with multiple legal jurisdictions around the world.  

Below is a chart showing the breakdown of the sexual activity seen in this 3–6-year-old 'self-generated' material.  

We can see that: 

  • 60% shows sexual posing with nudity;  
  • 16% shows rubbing of genitals (categorised as masturbation);  
  • 14% shows penetration. 

Sexual activity metadata: 'Self-generated' and 3-6-years-old

Sexual posing with nudity was the activity seen the most by far. The typical behaviour seen in this type of image was of children doing the following:  

  • pulling down underwear to ‘flash’ their genitals;  
  • dropping their underwear to their knees or ankles and opening their legs to show their genitals to the camera; 
  • bending right over with their bare bottoms towards the camera, showing their anus and genitals to the camera; 
  • Using their hands to pull apart their buttocks and or genitals to display themselves. 

Most of the time these children are initially clothed and much of what we see is a quick display of genitals. This could indicate the opportunistic nature of the perpetrators. It could also be that most 3–6-year-olds are not left alone long enough for the discussion and the coercion to get further along, towards full nudity and more severe sexual activity.  


Multiple children analysis

We looked at how many of these images contained multiple children. 

  • 28% (681) of the 2,401 images or videos showed a 3–6-year-old with at least one other child.   

Where multiple children were seen in the images and videos, we saw that Category C images accounted for nearly half. In these images the children are often displaying their genitals and are with another child who may or may not also be displaying themselves. They are not directly interacting sexually or touching each other.  

However, there was also a higher percentage of Category B images that had more than one child. Category B images include those where a child is rubbing genitals (categorised as masturbation) or where there is non-penetrative sexual activity which is where the children are interacting, perhaps touching each other in a sexual manner. 

  • 46% were Category C. 
  • 34% were Category B. 

Severity: Multiple children, 'Self-generated' and 3-6-years-old

  • 73% of the children in the multiple child images were girls. 
  • 22% of the images showed both boys and girls. 

Sex of children: Multiple children, 'Self-generated' and 3-6-years-old

  • Girls - 500 (73%)
  • Boys - 29 (4%)
  • Both - 152 (22%)

When there were multiple children in the imagery: 

  • Sexual posing was the activity seen most (43%)  
  • Non-penetrative activity was the next highest (23%), perhaps reflecting the fact that there were multiple children in the images, and they were interacting with each other in a sexual manner.  

Sexual activity metadata: Multiple children, 'Self-generated' and 3-6-years-old


Visual analysis of the images

We took a closer look at a small sample of these images to further investigate the activity seen. We sampled 202 images and videos; 130 images were of a single child and 72 contained multiple children.  

Most of the images and videos showed children in a home setting and most often in a child’s bedroom. In the backgrounds, analysts saw soft toys, games, books and bedding featuring cartoon characters. In some images the location could not be determined as the image was very close up.   

Full details of the locations seen in this sample group of images are below.  

Location Number of images and videos  %





58 29% 


24 12%

Living room 

18 9% 


6 3%

Dining room 

2 1%





Individual children

Category  Number of images and videos  % 
A 8 6%
B 30 23%
C 92 71%
Total 130 100%


Multiple children

Category  Number of images and videos 
A 5 7%
B 25 35%
C 42 58%
Total 72 100%

Category C was the grade given to the majority of the images with a slightly higher proportion of Category B among the multiple child images which also reflects the full data for the year.  

We looked further into the ages of the other children in the multiple child images and found the following: 

  • 72 images showed a total of 150 children.  
  • 66 images showed 2 children. 
  • 6 images showed 3 children.  
  • 79% (57) showed a 3–6-year-old with at least one older child. 
  • 28% (20) showed children of the same 3–6-year-old age group. 
  • Within the 72 images featuring multiple children:  
    • 13 boys were seen
    • Of those images featuring boys, 4 were also seen with girls and 9 boys were in images with other boys.  

It is notable that when multiple children were seen, the 3–6-year-olds were most often seen with an older child, perhaps a sibling. 

  • 58% of multiple child images were Category C and showed a child or children displaying their genitals and posing sexually.  
  • 35% of the multiple child images were category B and showed non-penetrative interaction between the children, for example sexual touching or licking.   

Below is the breakdown of the sexual activity seen in the whole sample alongside the activity of those that showed multiple children.  


Sexual activity metadata: 'Self-generated' and 3-6-year-old

Rotate your device



All 'self-generated' child sexual abuse imagery is horrific, and our analysts sadly see it every day, but seeing so many very young children in these images and videos is particularly distressing. The images seen can range from a display of genitals to a child penetrating themselves or another child and all for the gratification of an unknown predator.   

It is perhaps surprising that there is not a higher ratio of multiple child images in the 'self-generated' 3-6 age group. It would be easy to assume that a child of that age would only engage in this type of activity on camera with the encouragement in person of an older child, leading the way, but shockingly this is not what we have seen. It also goes to show how successful the abusers are at manipulating very young children into sexual behaviour that the child is unlikely to have previously been aware of. It also demonstrates the dangers of allowing a young child unsupervised access to an internet enabled device with a camera.  

These are very young children, supposedly in the safety of their own bedrooms, very likely unaware that the activities they are being coerced into doing are being recorded and saved and ultimately shared multiple times on the internet.