From compliance to community: the 2025 Trust & Safety Summit..
trust and safety,
trust & safety summit,
t&s,
Published on Thu May 15 2025
Updated on Fri Aug 08 2025
7 minute read
Over two days in London, hundreds of trust & safety professionals, policymakers, engineers, and product leaders gathered to tackle some of the most pressing challenges facing digital platforms today. The 2025 Trust & Safety Summit was more than just an industry check-in—it was a call to reimagine how we protect users, build communities, and adapt to an increasingly complex digital world. Here are my key takeaways from two days of incredible conversations.
Is the future of moderation user-centric and modular?
Is it feasible to manage Trust and Safety for a user base exceeding 30 million with a team of only 20, including 16 engineers? Bluesky is attempting to do so.
Bluesky’s “composable moderation” model stood out as a radical rethink of how platforms can approach content control. Rather than enforcing a one-size-fits-all policy, Bluesky lets users customize what they see—choosing how (or if) to filter different types of content through opt-in labels or third-party filters.
“We accepted failure as a starting point and handed moderation capabilities over to our users.” — Aaron Rodericks, BlueskyIt’s a model that acknowledges different communities have different values—and it may become more relevant as platforms decentralize and users demand greater agency.
Proactive policy beats reactive scrambling.
TikTok’s Siobhan Oat-Judge shared how her team is iterating on policies in real-time to address new challenges like AI-generated content and synthetic impersonation. Rather than waiting for regulation to force action, they’re proactively defining boundaries to protect users.
“We interrogated all our policies to assess what adjustments were needed in order to cover new use cases.” — Siobhan Oat-Judge, TikTokThis mindset—of being ahead of the curve, not just compliant—is becoming the standard for responsible platforms.
Safety by design works.
Gaming leaders from EA, Epic, and Microsoft emphasized that when safety is embedded into product development, it just works—and often goes unnoticed because users are simply having a better experience.
From default settings for younger players to risk assessments during feature planning, “safety by design” is becoming a benchmark of product excellence. It’s not just about preventing harm—it’s about designing systems that enable positive interactions from the start.
AI is powerful - but not perfect.
AI is rapidly transforming content moderation, helping scale efforts across text, voice, and video. But as panelists warned, it’s not a silver bullet. Bias, context gaps, and evasion tactics remain persistent challenges.
“AI will augment, not replace, the trust & safety workforce.”The consensus? Use AI to support human moderators, not sideline them—and always keep evaluating its fairness and effectiveness.
T&S is a business imperative, not a cost center.
Again and again, speakers returned to one message: trust & safety drives business growth. A well-moderated platform retains users, boosts engagement, and protects brand reputation.
“Moderation should be considered a growth channel, not a cost.” — Nate Sawatzky, SupercellTrust isn’t just nice to have—it’s essential infrastructure. More companies are learning that safe spaces foster loyal communities and long-term success. The example shared by Kim and Natasha from Activision/ Microsoft where they’ve seen a 30% increase in user retention after proactive voice moderation was activated, speaks volume.
Silos kill strategy - cross-functional T&S is key.
The example shared by Holly Shepherdson, where a platform update in isolation, without consulting the T&S team, caused user reports to jump from "3,000 a month to 3,000 a day" powerfully demonstrates the consequences of departmental isolation. Whether it’s policy, enforcement, engineering, or legal, trust & safety only works when these groups are aligned. Breaking silos means earlier intervention, smoother execution, and a stronger culture of safety.
Getting everyone—from product to legal—rowing in the same direction is the next frontier of organizational trust & safety maturity. To boost collaboration, tailor communication to each team's perspective and establish a network of "anchors" to embed shared principles across the organization. Prioritize presenting information, especially regarding risks, in a way that aligns with other teams' priorities, aiming to be seen as a partner rather than an obstacle.
The power of shared knowledge in a safer online world.
Yoel Roth, VP of Trust & Safety at Match Group, strongly emphasized the critical value of open knowledge and experience sharing within the Trust and Safety community. He highlighted that, given the field's constant evolution and the emergence of novel challenges, fostering a culture of transparent communication and the dissemination of learned insights through platforms like blog posts and other publications is essential for collective growth. Historically, many companies have been understandably cautious about divulging details of their Trust and Safety practices, concerned about potential risks. However, there's a growing recognition that a more open and communicative approach is not only beneficial, but crucial. Such transparency fosters greater trust with users and facilitates constructive feedback, even when it includes criticism. Roth encouraged practitioners to present preliminary versions of their communication strategies and actively invite collaboration from their peers, soliciting assistance in determining the most effective path forward.
He specifically encouraged those in T&S roles in organizations that aren't social media companies to submit TrustCon proposals, publish about what they’re doing, write blog posts, and share the knowledge and experience they have. While social media companies have accumulated 15 years of learning about what works, it's time for the rest of the industry to step up.
Final thought: the bar is rising.
What was once optional is now expected. Users, regulators, and stakeholders are holding platforms to higher standards—and rewarding those who lead with trust.
If we want safer digital spaces tomorrow, the work starts today. Not just by reacting to harm, but by building for safety from the beginning—collaboratively, proactively, and with heart.
Want more insights or to collaborate on future T&S research? Let’s connect. The more we share, the stronger our collective impact.
Read more
Created at Thu Sep 17 2026
6 min read
Leading Voices: most brands fail at agentic AI. Our solutions experts share how yours can win.
Something happens when a technology concept outpaces the terminology we have to describe it. Across the customer experience landscape, vendors start applying the same label to wildly different things. For instance, industry leaders start buying into the AI label instead of its real operational capability. And somewhere between the pitch deck and the production environment, the real question gets lost: does any of this actually address a real operational need? Agentic AI has reached that stage.
