Share icon
Sign up for our newsletter
Close icon

    Enter your email below to receive a free e-newsletter with the latest IWF news, industry and parliamentary updates and events direct to your inbox.

    Machine learning with SafeToNet

    We’ve been working with safety tech company SafeToNet to develop a new online child safety solution which will use machine learning to recognise child sexual abuse material and stop it being sent to, or created by, mobile devices.

    The solution, called SafeToWatch, will detect threats in real time, block harmful content or switch off the camera on a child’s phone if they are viewing inappropriate material or tricked into taking sexual imagery of themselves by an online predator.

    In 2022, 78% of the websites we removed contained images or videos where children had been groomed and coerced into sexual activities over an internet-enabled device with a camera by an abuser.

    We hope this new technology will help prevent children from being targeted in this way by criminals.

    Our tech team helped train SafeToWatch, using known child sexual abuse material assessed by our experienced IWF analysts, to teach it what child sexual abuse material is and how to detect it, and then rigorously tested it.

    SafeToWatch can sit on the device itself and will work, even in end-to-end encrypted platforms, to help keep children safe. It is also being developed to be integrated on platforms themselves and could be an important child safety solution as more platforms fully encrypt their services.

    “It is prevention at the point of creation or distribution. We are unique in that we are focusing on a preventative solution rather than a reporting tool. The reason we wanted to work with the IWF is because we believe they have the best and richest data worldwide. It is well labelled and lends itself to helping with this kind of work.” – Tom Farrell, Chief Operating Officer at SafeToNet.