For journalists & media partners
Boilerplate, brand assets, founder background, and a direct line for press inquiries. Use anything here without asking, but we'd love to hear from you anyway.
About Polycreek (Boilerplate)
Copy-paste paragraphs for news articles, press releases, and profile pieces.
Polycreek is a Denver-based 501(c)(3) nonprofit building AI tools to detect and prevent online child exploitation. Their AI model, Aletheia, identifies grooming behavior and abusive content in real time.
Polycreek is a Denver–based 501(c)(3) nonprofit developing AI-driven technology to protect children from online exploitation. Founded in 2024 by Zeran Johannsen, it operates at the intersection of machine learning and trust & safety. Polycreek integrates with systems such as Microsoft PhotoDNA and the NCMEC CyberTipline to support automatic reporting workflows. Its flagship AI model, Aletheia, analyzes digital conversations to identify grooming patterns and generate reports for review and escalation. The organization emphasizes rapid detection, responsible data handling, and collaboration with platforms and law enforcement to strengthen early intervention and prevention.
Online child exploitation is growing in scale and complexity, with millions of reports submitted annually to organizations like NCMEC. While existing tools are effective at detecting known illegal images, they are less equipped to identify grooming: gradual, conversational manipulation that often precedes abuse and leaves little immediate evidence.
Polycreek is a 501(c)(3) nonprofit focused on addressing this gap. It develops machine learning models trained to recognize linguistic and behavioral indicators of predatory contact. Its AI model, Aletheia, is designed to integrate with messaging systems across social, gaming, and community platforms, where it analyzes conversations in near real time to flag potential grooming behavior for human review.
Polycreek prioritizes early detection, privacy-conscious design, and interoperability with existing reporting pipelines, aiming to support faster intervention while minimizing unnecessary data retention.
Story angles we can speak to
Frames and angles where Polycreek can offer informed commentary, data, and on-the-record perspective.
The “Prevention vs. Prosecution” Shift
Most systems activate after harm occurs. The goal is moving intervention upstream: detecting risk before abuse escalates, and what that means for outcomes, ethics, and policy.
The Human-in-the-Loop Imperative
AI doesn’t replace investigators or moderators, it prioritizes and augments them. This angle focuses on reducing reviewer burnout, improving triage quality, and avoiding over-reliance on automation.
Privacy-Safe Detection: Can You Do Both?
A big tension in this space: detecting harm without expanding surveillance. We want to build something that has minimal retention, no unnecessary storage, scoped analysis and differentiate from more invasive approaches.
The Infrastructure Gap for Smaller Platforms
Large tech companies have trust & safety teams; smaller platforms don’t. Polycreek wants to enable small and mid-size platforms, gaming communities, and startups to access protections that were previously out of reach.
The “Invisible Harm” Problem
Grooming often leaves no artifact until it’s too late. This frames the issue as one of visibility. This is why existing metrics undercount harm and how conversational signals change that.
Globalization of Risk
Online exploitation isn’t geographically bounded, but laws and enforcement are. This opens up conversations about cross-border coordination and uneven protections.
Founder
Zeran Johannsen
Founder & President, Polycreek
Zeran Johannsen is the founder of Polycreek, a 501(c)(3) nonprofit dedicated to preventing online child exploitation through the responsible application of machine learning. He established the organization in 2024 to address critical gaps in existing safety infrastructure, particularly the early detection of grooming behavior in digital environments.
With a background in software engineering and machine learning, Johannsen focuses on translating emerging AI capabilities into practical tools for trust & safety. His work emphasizes measurable impact, cross-sector collaboration, and privacy-conscious system design. Through Polycreek’s proprietary model, Aletheia, he aims to support earlier intervention and strengthen reporting pathways for platforms, nonprofits, and law enforcement partners.
Brand Assets
Logo variations and brand colors. All logos are SVG.
