Social media platforms are optimized for outrage, not understanding. By rewarding extreme content with outsized reach, platform algorithms have accelerated the breakdown of healthy online norms—resulting in:

Escalating polarization and partisan hostility | Dehumanization of ideological opponents | Outrage loops that drown out nuance | Erosion of interpersonal trust | Disengagement from civil discourse and community life

Efforts like moderation, fact-checking, and labeling often backfire, reinforcing division or feeding claims of censorship. Meanwhile, those who want to model better online behavior struggle to keep up.

Normsy fights fire with (prosocial) fire. We meet harmful content where it spreads—on public threads, in real time—and respond not with censorship, but with normative counterspeech. That means modeling the social norms that guide healthy, respectful disagreement in real life but often vanish online.

Our approach rests on three principles: 1. Trust is relational—you can’t rebuild institutional or civic trust without first rebuilding interpersonal trust. 2. Norms shape behavior—what people see as acceptable online becomes what they do. 3. Technology can help—when paired with human insight, AI can scale constructive engagement, not just amplify division.

Key Features: Real-time detection of harmful content: Normsy monitors public social media conversations to identify toxic or polarizing posts as they gain traction—focusing on incivility, dehumanization, anti-democratic rhetoric, and more. AI-assisted, human-powered responses: A blend of generative AI and trained human responders deliver replies designed to de-escalate tension, model healthier behavior, and shift the tone of conversations. Evidence-based response strategies: Normsy draws from six social science-backed techniques—like showing empathy, highlighting shared values, and respectfully exposing inconsistencies—to guide constructive engagement. Simple, scalable intervention tools: Normsy’s dashboard allows users to generate thoughtful, context-aware replies with just a few clicks. No special training or deep expertise required. Built-in testing and optimization: Every response type is continuously evaluated to track what works, for whom, and under what conditions—ensuring ongoing learning and effectiveness.

Social media is shaping how we talk to each other—and how much we trust each other. Right now, the loudest voices online are often the most extreme. Left unchecked, this accelerates division, weakens community bonds, and erodes trust in everything from neighbors to institutions. But here’s the good news: Norms are contagious. Trust can be rebuilt. Normsy is a new kind of response—one that doesn’t censor, but constructively counters the toxic narratives dominating digital spaces. By modeling empathy, shared values, and respectful disagreement, Normsy helps shift the tone of entire conversations—for posters, for bystanders, and for the broader audience watching.

Our Goals: Reach 1 billion impressions by the end of 2025 by responding to toxic content with trust-building counterspeech. Deploy on 5+ major platforms by 2027, expanding from X to include Reddit, Threads, Bluesky, YouTube comments, and more. Drive measurable shifts in online behavior, reducing toxic norms and increasing exposure to prosocial engagement. Keep it accessible, maintaining a cost-per-impression lower than traditional sponsored content.

We believe the best way to fight polarization isn’t to silence voices—it’s to amplify better ones. With Normsy, we’re giving people the tools to do exactly that.