{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.text}}
Key takeaways:
- Community engagement needs to influence real product, policy, and enforcement decisions.
- Users are more likely to trust platforms when community feedback visibly shapes Trust and Safety policies.
- Engagement should be embedded into core workflows, not treated as a standalone activity.
- Leaders need to measure and communicate what changed as a result of community input.
For years, online platforms have invested in surveys, advisory councils, and stakeholder roundtables to demonstrate they’re listening to community feedback. These efforts generate visibility, but too often they fail to influence how decisions are actually made. Users expect a real voice in how platforms operate. Advertisers want the values to reflect their brands, regulators are increasingly scrutinizing governance and accountability, and AI systems are taking on a greater role in shaping user experiences.
PwC’s Trust and Safety Outlook 2026 research bears this out: 67% say they would be more likely to use a platform that regularly consulted its user community to guide policy and product design, and 72% say they would trust a platform more if it visibly incorporated community feedback into its Trust and Safety policies and decisions.
When the community sees that input does not actually shape policy, product design, or enforcement, platforms risk their trust efforts being perceived as performative rather than meaningful. As that perception gap widens, it becomes harder to strengthen trust.
The next phase of Trust and Safety is embedding engagement into the processes where decisions are made. It also requires being able to show what changed as a result.
Why current engagement models fall short
When asked how often the platforms they use most reflect their community’s values, only 23% of respondents say “always”, and 43% say “only sometimes, rarely, or never.” This is a sizable number of people who don’t consistently feel their values are represented.
Community input on Trust and Safety is often collected through consultations, surveys, or advisory groups, but is rarely integrated into the workflows that govern policy, product, or enforcement.
As a result:
- Communities don’t see impact, reducing trust and sustained participation. US users say they’d be motivated to take part in platform governance, but feel their input won’t make a difference, specifically around transparency and accountability (42%), and the chance to meaningfully influence platform decisions (33%). This is what’s missing when input doesn’t visibly shape outcomes.
- Internal teams struggle to operationalize input into products, policies, and operations.
- External stakeholders perceive engagement efforts as symbolic rather than substantive.
Even when input does influence decisions, the impact is not consistently tracked or communicated. Without clear evidence of what changed, engagement efforts remain difficult to validate.
What leading engagement models look like
AI is enabling organizations to make decisions faster, at greater scale, and with less direct human visibility. Relying on static policies or one-time input is not enough. Platforms should use mechanisms that continuously translate community expectations into how systems behave in practice, along with ways to measure whether those mechanisms are working.
Leading platforms are moving toward models where community input directly informs decisions. Several are already operationalizing this shift in distinct ways.
Deliberative engagement—structured input that informs decisions
Platforms are experimenting with more structured forms of engagement that move beyond open-ended feedback. Meta’s Community Forums, run in partnership with Stanford’s Deliberative Democracy Lab, convene representative samples of users, such as a gathering of 6,400 people across 32 countries for the Metaverse forum, who review briefing materials and debate specific policy proposals in small, facilitated groups.
After that forum, Meta added directly informed Meta’s decision to add mute assist, an automatic speech-detection tool, to the creator toolkit in Horizon Worlds. This begins to close the gap between public input and product decisions, something traditional engagement models rarely achieve.
Moderator ecosystems—distributed enforcement capacity
On community-driven platforms, moderators are a core part of how rules are enforced. Platforms like Discord are investing in training, tooling, and support to make moderation more consistent, directly shaping how policies are applied at scale.
This model allows platforms to scale enforcement through communities, while also introducing new expectations around consistency, support, and accountability. Because moderators operate within enforcement workflows, their impact is more observable, from escalation patterns to enforcement outcomes.
Community alignment—embedding values into system behavior
Meta’s Community Alignment program, released as an open-source data set in 2025, highlights a broader Trust and Safety challenge: how platforms can design governance systems that reflect the values of a global user base whose preferences may conflict across cultural, political, and linguistic lines.
The program found that people show substantially more variation in their preferences than leading LLMs typically reflect, suggesting that standard alignment methods can unintentionally reinforce an average user model that obscures underserved or underrepresented perspectives. The data set is designed to preserve that variation by having annotators compare model responses across known dimensions of value difference, with multiple people rating the same prompts and providing natural-language explanations for their choices.
For Trust and Safety leaders, the implication extends beyond AI alignment alone: Governance should measure whether product design, policy development, enforcement operations, and automated systems reflect the full distribution of user expectations, preserve meaningful cross-cultural differences where appropriate, and continue to stay aligned as platforms, policies, and user communities evolve over time.
How Trust and Safety leaders should think about engagement
Moving beyond symbolic engagement requires treating engagement as a core governance capability and element of product and policy development, rather than a standalone activity. Leaders should focus on four priorities:
1. Start with the decision you want to change
Define up front what product, policy, or operational decision engagement is meant to influence. Without a clear target, engagement efforts risk generating input that never translates into action.
2. Tie engagement to decision-making
Clarify where and how community input directly influences policy, product, or operational designs—and where it does not. This is what users value most: the opportunity to meaningfully influence platform decisions ranked highest, with a third (33%) of respondents naming it as a top reason to participate. Engagement that visibly leads nowhere erodes the very willingness to engage.
3. Embed engagement into core workflows
Facilitate processes that allow user input to feed into real workflows, such as product design, policy development, and operating model definition, rather than existing as a parallel activity.
4. Measure and communicate impact
Leaders should define how engagement will be measured up front, not just in terms of participation, but in terms of outcomes. Those outcomes may include changes to products, policies, and enforcement practices, improvements in operational performance, and stronger trust among affected communities.
Platforms should also communicate what changed as a result of engagement, whether through transparency reports, product updates or direct feedback loops with participants. The demand for visibility is clear: 80% of US users say it’s important that platforms explain how community feedback influences Trust and Safety policies and enforcement decisions, with 42% calling it extremely important.
Looking ahead
Platforms that lead in Trust and Safety can be better positioned to show how community input informs decisions, drives meaningful outcomes, and helps build both trust and better user experiences.
That doesn’t mean every piece of input will change a decision—nor should it. The more important work is being transparent about what was heard, what changed, what didn't, and why. Platforms that can do this consistently can earn sustained participation. Those that can’t will find internal and external engagement efforts increasingly difficult to justify.
Trust and Safety Outlook 2026
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}
FAQs
Follow us
