Introducing our chatbot - a pioneering tool in the fight against online child sexual abuse

The National Crime Agency estimates that there are up to 830,000 adults in the UK that pose some degree of sexual risk to children, and last year the Internet Watch Foundation found more than 275,652 web pages containing images and videos of child sexual abuse. The impact on the children abused is lifelong, and their suffering is compounded every time someone shares or accesses these images and videos. We know from our work, too, that the impact on the families of people arrested for offending is enormous.

Each year the numbers climb. And as they do, we continue to search for innovative ways to prevent this harm – with resources such as our chatbot. This pioneering online tool is designed to confront and redirect people searching for sexual images of children. Already operational on Pornhub UK, the chatbot has shown significant success in reducing searches for illegal material and directing people to our Stop It Now helpline and online self-help. The chatbot is the first of its kind in the world.

How our chatbot works

The chatbot is both a deterrent and a resource. When someone attempts to search for illegal content, the chatbot intervenes, issuing a warning and providing information about the potential consequences of their actions. Crucially, it also guides these people towards our help and support to change their behaviour. Working with people who might harm is one of our key objectives, as it helps to protect children by taking potential offenders away from their pathway and towards better choices.

Early success and impact

The chatbot’s impact is central to our latest Faithfull Paper:‘Chatbots and Warning Messages - Innovations in the Fight Against Online Child Sexual Abuse’. The paper offers an in-depth analysis of the tool and places it within the broader context of our other deterrence efforts, including the deterrence campaign and our tech project, Intercept. It also looks at the origins of the chatbot, which began life as a collaboration between us, the Internet Watch Foundation and Aylo, the owners of Pornhub. It was important to us to place the chatbot in a context where we could confront large numbers of people looking for child sexual abuse images and videos, and while Pornhub maintains a zero-tolerance policy for this illegal material, people continue to search for it regularly on the site, providing a valuable opportunity to test the chatbot and see how it might work.

An evaluation report on the reThink chatbot was published earlier this year which revealed promising results.

• Reduction in searches: the presence of the chatbot on Pornhub UK has led to a statistically significant decrease in searches for sexual videos of children.
• High deterrence rate: impressively, just one warning from the chatbot was enough to stop people in 82% of sessions (periods of time spent on the website) from continuing their search for illegal content.

These findings highlight the potential of technology in combating online child sexual abuse, and the importance of early intervention.

Integration with Project Intercept

The chatbot is coming to sit within Project Intercept, which works with tech companies to create timely, impactful warnings to people at risk of offending. We already work with a wide range of platforms and tech companies who welcome the warnings as they are effective, easily implemented and cheap. In addition, they work at the secondary level of prevention – that is, targeted at people who are at risk of offending, but before harm has happened.

Sarah Smith, the lead for Project Intercept, said, “The more tech companies who sign up to work with us, the wider our reach with these messages will be, and the more chance we will have to intercept people who are at the point of offending. Collaboration between the online tech sector and the third sector is key if we are to tackle child sexual abuse at scale.”

We know from our work with people who have offended that in many cases a well-placed warning can be enough to nudge them back to legal behaviour. One of Project Intercept’s aims is to make warning messages the norm rather than the exception: what might things look like if every website had a warning message or a chatbot to intercept those looking to harm children or view abuse material? Why shouldn’t potential offenders be challenged, deterred, reminded, nudged and redirected? And why shouldn’t tech companies be challenged to get on board and help make obvious changes that protect children?

Support from Ofcom and the Online Safety Act

We are pleased that our efforts are reinforced by regulatory support. Ofcom, the UK’s online safety regulator, recommends the use of warnings and support resources in its draft codes for the Online Safety Act. Although the current codes are aimed at large search services, this endorsement highlights the importance of proactive measures and the effectiveness of warning systems in preventing online abuse.

Future developments

We are exploring future possibilities for developing our chatbot.

• A generative AI-powered upgrade: can we make our chatbot smarter? The original chatbot was programmed with pre-written responses and was limited in the scope of the help it could offer. Could we develop a smarter, more responsive chatbot that drives more people to seek help?
• Expansion to other adult websites: building on our success with Pornhub UK, we would like to be able to deploy the chatbot across a range of adult websites to broaden its reach.
• Adaptation to other platforms: illegal content is sought and shared on a wider range of sites and platforms than might be expected, from music and file-sharing sites to dating, social media and gaming. We are exploring ways to make the chatbot adaptable for use on as many digital platforms as possible.
• Multilingual support: to extend our impact globally, we plan to develop versions of the chatbot in multiple languages.
• Targeting resistant users: the chatbot evaluation found that a very small subsection of users ignored our warning messages and continued to search for illegal videos of children. Future chatbot experiments could target this group.

Join us in the fight against child sexual abuse

Child sexual abuse is preventable, not inevitable. The new chatbot is a key part of our ongoing efforts to leverage technology in the fight against this abuse. We invite you to stay informed, get involved, and support us as we continue to develop and expand our initiatives.

Find out more

If you would like more information about the chatbot or Project Intercept, please contact: intercept@lucyfaithfull.org.uk

Read our Faithfull Paper about the impact of our campaign to deter people from online child sexual abuse.

Want to support our work to prevent child sexual abuse? Donate now.

Sharing our posts and information helps us reach more people and protect more children. Sign up to receive our emails and find us on Twitter/X, Facebook, LinkedIn, Instagram and YouTube.