🎥 From the MLOps World | GenAI Summit 2025 — Virtual Session (October 7, 2025) Session Title: The Efficiency Equation: Leveraging AI Agents to Augment Human Labelers in Building Trust & Safety Systems at Scale Speaker: Madhu Ramanathan, Senior Engineering Leader, Trust and Safety, Meta Talk Track: Augmenting Workforces with Agents Abstract: In today’s fast-moving digital ecosystem, Trust & Safety systems must keep pace with content proliferation, real-time enforcement demands, and ever-changing threats like deepfakes, hallucinations, misinformation, and adversarial behavior. In this keynote, Madhu Ramanathan explores how LLMs and AI agents are transforming Trust & Safety operations by augmenting human labelers—delivering scalable, high-quality, and cost-efficient systems that balance defect rate, precision, and cost. The session dives deep into: • The evolution of Trust & Safety systems and emerging AI-driven challenges • The traditional role of human labelers across enforcement, measurement, and reactive takedowns • Real-world case studies of how LLMs and agents are used to measure, detect, and enforce at scale • Continuous evaluation and calibration techniques using human-in-the-loop and automated prompt tuning • Solutions to complex production challenges like model migration, GPU cost constraints, and bias mitigation Through Meta’s experience, Madhu reveals how human judgment and AI intelligence can co-evolve to create hybrid Trust & Safety systems that are resilient, ethical, and continuously improving. What you’ll learn: • Why Trust & Safety is entering a new AI-augmented era • How hybrid systems balance human oversight with agentic automation • How to design resilient, cost-effective, and adaptive enforcement systems • The future of combining human intelligence (HI) and artificial intelligence (AI) for scalable trust solutions.