Visual Report and Launch Event
A co-created visual report exploring how social scientists and computer scientists can use AI to combat online gender-based violence, led by young people.
Developed through a collaboration between University of Edinburgh and Heriot‑Watt University, Equally Safe Online places young people’s voices at its centre, shaping not only the research, but how it is communicated and shared.
The work documents the development of an AI chatbot designed to generate counter-speech to online abuse, showing how technology can be used for good when it is guided by lived experience.
University of Edinburgh · Heriot-Watt University · With involvement from Google DeepMind and youth participants
Visual report design · Editorial illustration · Participatory Practice · Exhibition graphics · Motion/video
2025
The Challenge
Online gender-based violence is widespread, harmful, and often invisible in traditional research outputs.
The challenge was not only to address the proliferation of hate speech online, but to communicate a hopeful alternative: that AI, when shaped with care and accountability, can help generate meaningful counter-speech and in doing so can counteract the negative effects of hate speech.
A design-led approach was essential to ensure the research did not remain abstract or inaccessible. The work needed to speak clearly to policymakers, technologists, and funders while remaining grounded in the voices of the young people most affected.
Our Approach
Young people’s input was centred from the very beginning of the project.
We created an early eight-page draft of the report, visually rich and graphic-novel-like, and invited young participants to respond directly to both the content and the way it was being communicated.
Tone was critical. The subject matter is serious and sometimes painful, so the visual language needed to be approachable, respectful, and engaging, without ever softening the reality of what was being discussed.
Shaping the report. Photograph by Elspeth Maxwell
The Outcome
Together, these elements created a cohesive visual language that could move fluidly between research, public engagement, and policy contexts.