Accurately moderating age-related content is essential for platform safety—but AI models often misclassify casual age mentions, overwhelming moderation queues. This session presents a two-layered AI system deployed at a large social platform with millions monthly active users. The first multilingual classifier flags potential underage content; a second, context-aware filter analyzes behavior patterns to refine decisions. This approach reduced false positives, accelerated sanctions, and improved moderation efficiency. Attendees will learn how to build scalable workflows, leverage behavioral signals for precise filtering, and elevate compliance—all while boosting user trust and operational effectiveness.

Let’s explore the multifaceted nature of chargeback fraud—examining its detrimental impact on both marketplaces and the financial institutions tied to compromised credit cards, along with strategies to effectively mitigate these risks.

We all know that background checks are essential to risk mitigation. But if not conducted clearly you could create more risk to your company with litigation and regulatory enforcement action. Learn about considerations for background checks generally, specifics in New York City and California as it relates to marketplaces.

Platforms are ecosystems that sit within a greater online ecosystem. This presentation discusses why and how adopting an ecological mindset that acknowledges the porosity of online spaces can drive more effective platform policing strategies.

Fraud never stops; it constantly reinvents itself. From device resets and GPS spoofing to behavioral mimicry and synthetic identities, bad actors are evolving faster than ever. Modern fraud prevention goes beyond technology, encompassing product design, policy, and user experience. This panel brings together fraud strategy and enforcement leaders to explore how platforms are advancing their defenses by focusing not just on who is returning, but how. Through real-world investigations and detection frameworks, the discussion reveals how modern signals like location behavior and device intelligence help identify repeat fraud that traditional rules miss. Key takeaways include detecting reoffenders after resets and spoofing, insights from investigations into seller abuse and identity evasion, operationalizing new signals to outpace attackers, and the importance of adaptable, privacy-conscious detection strategies to preserve user trust.