Adopting a user-centred design approach for the development of on-device technology to prevent the viewing of child sexual abuse material: app design insights and principles from the development of 'Salus'.

Introduction: The volume of Child Sexual Abuse Material (CSAM) available online and the global demand for it has reached unprecedented levels. Increasing numbers of individuals concerned about their online behaviour are contacting therapeutic providers for help and support outside of the criminal justice system. Previous research asking individuals what would help them to stop viewing CSAM suggests that the availability of a technological solution to voluntarily self-manage access to CSAM could be an effective tool. Aim: To explore the findings from the user-centered design (UCD) of the 'Salus' prototype - a technological prevention tool to support effective self-management of individuals at risk of committing a first or further CSAM offence(s). Materials and methods: In this two-year, European Commission funded project we conducted research in four European countries: Belgium, Germany, the Netherlands, and the United Kingdom (UK). For the UCD phase of the project we conducted semi-structured interviews with 31 at-risk individuals in Belgium (n=10), Germany (n=10) and the UK (n=11), to explore the specific needs, design features, deployment methods, and concerns and barriers for the design, functionality and deployment of Salus. Additionally, four focus group discussions (FGDs) were held in Belgium, the Netherlands, and the UK with service providers (primarily therapists and managers) with extensive experience of supporting individuals at risk of committing CSAM offences to explore the same questions at the service level. Results: In terms of privacy and security, the potential discovery of apps such as Salus, data security and legal consequences of app usage are the main concerns of potential app users. There was consensus on the value of blocking CSAM, but opinions on the inclusion of an optional adult sexual content (pornography) filter in Salus design were not unanimous. Users should be able to switch a pornography filter on and off at their convenience. Blocking notifications should be quiet and subtle. Interactivity features are welcomed by potential users - these may include a diary function; a personal CSAM statistics page; a resources section; and a function to allow users to provide feedback to the app developers. Such features should be optional for users in order to prevent any unintended consequences of app usage. Finally, app deployment must be safe and secure. Conclusion: Based on these findings, we propose seven evidence-based design principles for user-centered harm-reduction technology: privacy-by-default architecture; discretion through design ambiguity; adaptive notification systems; optional interactivity with user control; trusted-channel deployment; progressive trust building; and fail-safe harm prevention. These principles provide a framework for app developers and researchers working on similar technologies to develop interventions that reduce harmful behaviours.