Session
Description:
Artificial intelligence is rapidly transforming how online platforms identify, assess, and respond to harmful content. While AI promises faster detection and greater scale, many online harms, including technology-facilitated gender-based violence (TFGBV), AI-generated intimate image abuse, coordinated harassment, and gendered disinformation, continue to require contextual understanding, local expertise, and human judgment. Drawing on insights from Digital Rights Foundation's Digital Security Helpline and the Feminist Helplines Research Initiative - a collaborative research initiative involving fourteen feminist helplines across the Global South - this workshop explores how civil society is already contributing to platform governance through survivor support, trusted reporting, evidence generation, policy engagement, and accountability mechanisms. Rather than debating whether AI should replace human moderators, the session asks a broader governance question: How can AI systems, trust and safety teams, civil society, governments, researchers, and users share responsibility for creating safer online environments? Through an interactive governance design exercise based on real-world anonymized cases, participants will examine how online harms move through today's trust and safety ecosystem, identify where existing systems succeed and fail, and collaboratively redesign governance processes that better integrate automated systems with human expertise. Bringing together representatives from technology companies, civil society organizations, researchers, and policymakers, the workshop will focus on practical governance solutions rather than abstract principles, identifying concrete ways to strengthen accountability, transparency, and collaboration in AI-enabled trust and safety systems.
Artificial intelligence is rapidly transforming how online platforms identify, assess, and respond to harmful content. While AI promises faster detection and greater scale, many online harms, including technology-facilitated gender-based violence (TFGBV), AI-generated intimate image abuse, coordinated harassment, and gendered disinformation, continue to require contextual understanding, local expertise, and human judgment. Drawing on insights from Digital Rights Foundation's Digital Security Helpline and the Feminist Helplines Research Initiative - a collaborative research initiative involving fourteen feminist helplines across the Global South - this workshop explores how civil society is already contributing to platform governance through survivor support, trusted reporting, evidence generation, policy engagement, and accountability mechanisms. Rather than debating whether AI should replace human moderators, the session asks a broader governance question: How can AI systems, trust and safety teams, civil society, governments, researchers, and users share responsibility for creating safer online environments? Through an interactive governance design exercise based on real-world anonymized cases, participants will examine how online harms move through today's trust and safety ecosystem, identify where existing systems succeed and fail, and collaboratively redesign governance processes that better integrate automated systems with human expertise. Bringing together representatives from technology companies, civil society organizations, researchers, and policymakers, the workshop will focus on practical governance solutions rather than abstract principles, identifying concrete ways to strengthen accountability, transparency, and collaboration in AI-enabled trust and safety systems.
Policy Question(s)
A. How should responsibilities for preventing, identifying, responding to, and learning from AI-enabled online harms be shared between technology companies, governments, civil society, researchers, and users?
B. How can frontline evidence generated by trusted reporting mechanisms, digital support services, and civil society organizations be integrated into the design, evaluation, and governance of AI-assisted trust and safety systems?
C. What governance mechanisms are needed to ensure that increasing reliance on AI strengthens—rather than weakens—human rights, transparency, accountability, and meaningful human oversight within platform trust and safety systems?
Expected Outcomes
Participants will co-create a ‘Shared Responsibility Framework for AI-enabled Trust & Safety’, mapping governance responsibilities across different actors and stages of the online harm lifecycle. The session will generate practical recommendations on integrating AI with human expertise, strengthening collaboration between platforms and civil society, and improving accountability within trust and safety systems.
The resulting framework and key recommendations will be compiled into a publicly available session summary for dissemination through participating organizations, contributing to ongoing multistakeholder discussions on AI governance, platform accountability, and online safety.
Format
Classroom
Duration (minutes): 90
Format description: The session is designed around collaborative problem-solving rather than presentations. Short opening interventions will establish common context before participants work together in small, multistakeholder groups using real-world governance scenarios. This format encourages active participation from all stakeholder groups while producing tangible outputs that can inform future policy discussions. A 90-minute workshop provides sufficient time for participants to engage meaningfully with the scenarios, develop governance recommendations, compare perspectives across stakeholder groups, and collectively identify areas of consensus and disagreement.
Duration (minutes): 90
Format description: The session is designed around collaborative problem-solving rather than presentations. Short opening interventions will establish common context before participants work together in small, multistakeholder groups using real-world governance scenarios. This format encourages active participation from all stakeholder groups while producing tangible outputs that can inform future policy discussions. A 90-minute workshop provides sufficient time for participants to engage meaningfully with the scenarios, develop governance recommendations, compare perspectives across stakeholder groups, and collectively identify areas of consensus and disagreement.
Hybrid Format: Online participants will participate in exactly the same exercises as onsite attendees. Using Mentimeter (or Slido), participants will first respond to live polling questions on responsibility for AI-enabled harms, establishing a baseline for discussion. Both onsite and online breakout groups will work simultaneously using shared digital collaboration boards (e.g., Miro) containing the Trust & Safety Journey Map and Shared Responsibility Matrix. Online facilitators will ensure virtual participants can contribute equally to group discussions. The final Shared Responsibility Framework will be compiled live using inputs from both onsite and online groups, ensuring that recommendations reflect the perspectives of all participants regardless of location.
Organizer 1: Nighat Dad, 🔒Digital Rights Foundation
Organizer 2: Maria Nazar, Digital Rights Foundation
Organizer 2: Maria Nazar, Digital Rights Foundation
Speaker 1: Nighat Dad, Civil Society, Asia-Pacific Group
Speaker 2: Alexandria Walden, Private Sector, Western European and Others Group (WEOG)
Speaker 3: Mohamad Najem, Civil Society, Asia-Pacific Group
Speaker 2: Alexandria Walden, Private Sector, Western European and Others Group (WEOG)
Speaker 3: Mohamad Najem, Civil Society, Asia-Pacific Group
Moderator
Maria Nazar, Civil Society, Asia-Pacific Group
Online Moderator
Hyra Basit, Civil Society, Asia-Pacific Group
Rapporteur
Hyra Basit, Civil Society, Asia-Pacific Group
What will participants gain from attending this session?
