AI for Good - Detecting Harmful Content at Scale

This talk examines how teams build and operate AI systems for detecting harmful content at scale. It covers testing, monitoring, model adaptation, deployment, and cultural sensitivity in content moderation.

Syllabus

[] Matar's preferred coffee
[] Takeaways
[] The talk that stood out
[] Online hate speech challenges
[] Evaluate harmful media API
[] Content moderation: AI models
[] Optimizing speed and accuracy
[] Cultural reference AI training
[] Functional Tests
[] Continuous adaptation of AI
[] AI detection concerns
[] Fine-Tuned vs Off-the-Shelf
[] Monitoring Transformer Model Hallucinations
[] Auditing process ensures accuracy
[] Testing strategies for ML
[] Modeling hate speech deployment
[] Improving production code quality
[] Finding balance in Moderation
[] Model's expertise: Cultural Sensitivity
[] Wrap up