Last week, researchers and industry partners gathered for the third year in a row – this time in Singapore – to bring together different perspectives and expertise for tackling online CSEA.
Hosted by Safe Online and the Tech Coalition on the margins of the Trust and Safety Professionals Association (TSPA) APAC Summit, this year’s Research Fund Convening brought together nearly 40 participants from across sectors and regions. The goal: to move from insight to implementation, to ensure that research doesn’t just describe problems, but helps shape solutions that work in the real world.
Opening remarks by Kay Chau from the Tech Coalition underscored the unique role that the Tech Coalition Safe Online Research Fund plays for industry:
“Since its inception, the Research Fund has supported nearly USD 3 million in independent studies, helping ensure that decisions within member companies are informed by evidence rather than assumptions.”
Snigdha Bhardwaj, Director of Trust and Safety at Google- who has been a great supporter of the Research Fund through hosting the convenings – kicked off the day with a discussion of their pioneering work with generative AI and how to proactively mitigate risks, emphasizing that partnership with other actors within the ecosystem is crucial to create a safer internet.
The first panel of the day brought in evidence from the region – the Philippines, Cambodia, Nepal, Kenya, and Australia – painting a clearer and more nuanced picture of how children and caregivers navigate their online lives.
Children have their own protective strategies that they employ in digital spaces. Many run quick “background checks” on new contacts, look for mutual friends, or test for consistency in conversation. They weigh benefits and harms almost instinctively—moment-to-moment calculations that adults often underestimate. And yet, the formal act of reporting remains challenging. Many children fear the process isn’t confidential or doubt that it leads to meaningful change. Others simply find that it breaks the natural rhythm of their digital experience – hinting at solutions by building reporting or help seeking tools into the regular flow of platform experiences.
The discussion turned from children to caregivers: the research from ChildSafeNet found that parents often know about screen-time limits but rarely about the “three Cs” of online safety: contact, content, and conduct. But the barrier isn’t just knowledge, but also dynamics. For instance, in many households represented in the research in Nepal, fathers possess stronger digital skills but tend to be more emotionally distant from day-to-day parenting. Mothers, meanwhile, often maintain closer emotional bonds but lack the same technical confidence.
Participants reflected on how parenting education could bridge that divide: combining digital literacy with emotional literacy, and helping caregivers recognise that protecting children online means being both technically capable and emotionally available.
Build research into the product cycle
Safety tools and interventions work best when they’re baked into how people already move online and responsive to the dynamics that shape their relationships. Understanding children’s protective strategies, and the realities of how caregivers support or fail to support them, can help designers and educators create features and programs that feel intuitive, empathetic, and effective rather than punitive or detached.
The next session peeled back the curtain on what collaboration between researchers and companies can really look like when it works. Several grantees shared experiences of designing and implementing studies in collaboration with tech platforms—from building advisory panels that inform researchers around aspects of approaches such as questionnaire wording for industry input, to distributing surveys that lead to insights that impact platforms’ design of products, policies or processes.
Most importantly, speakers were candid about the practicalities of collaboration: access to real-world systems, internal approvals, sensitivity reviews, and the need for legal agreements to protect competitive information. Some teams produced individualised reports for specific platforms under confidentiality, while publishing aggregated findings publicly.
Clear boundaries with industry partners, including being aligned on independent research questions, ensured that even when methodologies were adjusted (for example to be able to work with a platform’s internal approval processes and data-collection timeline), the scientific integrity of the results remained unimpacted.
These partnership models can yield results that neither side could achieve alone.
Partnerships necessitate trade-offs and structure, not just goodwill
Successful collaboration depends on clarity, trust, and mutual accountability. Building structure around partnership, through legal safeguards, shared objectives, and flexibility, turns potential friction into a pathway for impact.
Additionally, the convening featured all 11 grantees of the Research Fund presenting their work in a dynamic “lightning round” format. This rapid-fire session offered participants a quick but comprehensive overview of each research project. Following this, a “marketplace” session allowed for deeper engagement, giving attendees the opportunity to visit specific research teams, ask questions, and explore the projects in more detail through one-on-one and small-group discussions.
After lunch, the convening shifted gears from discussion to design. In mixed groups, researchers and industry participants explored how to make research both more inclusive and more actionable. The first working session on creative methodologies for inclusion surfaced inventive ways to involve children and young people, such as AI-supported participatory tools with the required, tailored ethical safeguards built into open models by child participation experts.
The second exploration focused on the enabling environment for translating evidence into practice. Participants shared the challenges of adapting research across geographies and user groups, and of aligning rigorous timelines with the pace of product development. Industry voices described the complexity of implementing findings within global platforms that must localize for distinct cultural and regulatory contexts. The consensus was that evidence translation is rarely linear: it requires flexibility, shared language, and mutual understanding of each sector’s constraints.
Strengthen evidence on what works
Participants emphasized the importance of systematically evaluating child online safety initiatives to identify which approaches are most effective and why. Sharing these findings – how interventions improve user experience and ultimately safety – can help all stakeholders make a stronger case for continued and scaled focus, effort and investment.
As the day drew to a close, participants shared one action they’d take forward. A number of key themes emerged:
One final discussion returned to diversity: both of technology and of childhood itself. The tech ecosystem isn’t a monolith, participants noted; it ranges from hosting services to social platforms, each with distinct leverage points for safety. And “children” aren’t a single audience either. Designing for those most vulnerable, such as children with disabilities, or those in low-connectivity settings, often produces solutions that benefit everyone.
Partnerships necessitate trade-offs and structure, not just goodwill
Successful collaboration depends on clarity, trust, and mutual accountability. Building structure around partnership, through legal safeguards, shared objectives, and flexibility, turns potential friction into a pathway for impact.
The 2025 Research Fund Convening reinforced the value of bringing researchers and industry together to bridge evidence and implementation. The conversations in Singapore will continue to inform how the Tech Coalition and Safe Online shape future priorities: advancing actionable, evidence-based approaches to keep children safe online.
🌐 Explore the new Research Fund webpage
The brand new Research Fund webpage, launched during the convening, now serves as a living record of impact — capturing key indicators, showcasing collaboration through an interactive heat map, and featuring case studies and best practices from grantees around the world. It offers an at-a-glance view of how evidence from the Fund is driving action across sectors and regions, and how partnerships are shaping a safer digital world for children.
Our grantees UNICEF Jordan Targeting Online Sexual Exploitation of Children in Jordan Countries involved:Jordan UNICEF Jordan is using support from the Fund to improve services for children and families experiencing online child sexual exploitation and abuse. To do so, the agency is building national capacity through the Prevention of Online Sexual
Safe Online’s latest funding round is investing $4 million for generating crucial data and evidence to tackle online child sexual exploitation and abuse (CSEA). This initial investment will be followed by an additional $1.5 million for evaluations of the projects in the next quarter.
Our grantees Middlesex University Invisible Risks Countries involved:United Kindgom of Great Britain and Northern Ireland We know that online child sexual abuse material is highly damaging to children. But today, little primary research exists about the impact such material has on content moderators – individuals who are charged with constantly
Subscribe to our monthly newsletter to keep up-to-date on the latest progress, news, events and reports from Safe Online and the wider community working to end violence against children.
Copyright Safe Online 2023 ©
All imagery is taken from the Safe Online and UNICEF image library
We are here to ensure every child and young person grows in to the digital world feeling safe, and is protected from harm.
We support, champion, and invest in innovative partners from the public, private, and third sectors working towards the same objective.
We believe in equipping guardians and young people with the skills to understand and see danger themselves once accessing digital experiences without supervision.
