Protect Children supports development of new ChatGPT intervention to prevent child sexual abuse material searches
Every child has the right to grow up safe from sexual abuse and exploitation. Stopping the demand for child sexual abuse material (CSAM) is one way to protect them. When people searching for CSAM are interrupted and directed to anonymous support, it can prevent harm to children before it happens.
We are glad to have supported OpenAI in developing a new evidence-informed intervention in ChatGPT that interrupts attempts to find CSAM and directs people to support. We contributed alongside Frances Frost (Lucy Faithfull Foundation), Joel Scanlan (CSAM Deterrence Centre), and Caoilte Ó Ciardha.
Our research shows why this kind of intervention matters. In Preventive Pathways, one of the largest real-world studies of CSAM warning messages, we found that every warning message we tested outperformed a neutral message, and that redesigning the messages nearly doubled the number of people clicking through to anonymous help.
In our survey of over six thousand people searching for CSAM, three in four of those who had seen a warning reflected on or changed their behaviour as a result.
Our CSAM Perpetrator Research Report, based on over twenty thousand survey responses, also shows how generative AI is reshaping CSAM perpetration, with 35% of respondents reporting that they had viewed or created AI-generated CSAM.
We welcome this step towards prevention that meets people early and moves further upstream.
Strengthening the fight against online child sexual abuse with new tools and support


