On Safer Internet Day 2020, INHOPE and the European Commission hosted a one-day specialist Focus Group on the potential development and implementation of Artificial Intelligence, incorporating Machine and Deep Learning to detect and remove online Child Sexual Abuse Material.

The day started with an overview of what AI is and what it is not and can never be, and what it can be used for today from AI expert Klaas Dijkstra. Participants of the panels and workshops that took place throughout the day included hotline representatives, technology and internet industry representatives, law enforcement, academia and cyber-psychology experts. We discussed what currently exists in terms of AI to combat online CSAM, what gaps there are and how this group of experts who have come together could work on furthering the potential of AI.

The two morning panel sessions addressed the realities of the new technology, as well as sharing the possibilities and current shortcomings of AI as a solution to remove online CSAM. Law enforcement colleagues presented the application of CSAM by the police and how it could be used more effectively and practically applied in the future.

In the afternoon there were two interactive workshops for the experts. One workshop was led by Jos Flury from ZiuZ, ICCAM’s technology lead for INHOPE, we discussed applied solutions for hotlines and LEA in integrating existing AI technologies. Much of the discussion centred around the need to define more specifically what we want AI to do, what the analysts and hotlines need AI to do and on solutions overcoming soft barriers to the application of AI. The other workshop was led by cyber-psychology and behavioural psychology expert Dr Mary Aiken who zoomed out to the bigger picture and the possibilities of AI. Recommendations included the need to invest in AI and the application of AI to work in a preventative manner, for example by stopping any online CSAM being seen more than once by analysts, or even never, as well as the need for self-regulation and the necessity of decisions made by AI to be revocable by humans.

If anyone would like to receive the full summary of the days' presentations and discussions, or if you have a question about this or future Focus Groups, please contact samantha.woolfe@inhope.org.