Me, Myself, and AI Protection, Provenance, and Prompts: YouTube’s Angela Nakalembe
12 snips
Sep 30, 2025 In this engaging discussion, Angela Nakalembe, an engineering program manager at YouTube and expert in internet safety, dives into the crucial role of AI in content moderation. She explains how AI serves as a first line of defense, improving response times and reducing harm to human reviewers. Angela also addresses the challenges of managing AI-generated content and the importance of provenance standards to combat misinformation. She shares her fascinating journey from chemical engineering to tech, alongside practical applications of AI in her personal life.
AI Snips
Chapters
Transcript
Episode notes
AI Protects Human Reviewers
- YouTube uses AI to act as a first line of defense, catching harmful content before human reviewers see it.
- This reduces emotional toll on reviewers and speeds removal of violative content.
Combine Models With Human Judgment
- Use AI to automate obvious policy violations and reserve humans for nuanced decisions.
- Combine models with human review to balance speed and contextual judgment.
Provenance Becomes Critical
- Rapid, frictionless AI content creation floods platforms with realistic but potentially misleading assets.
- This escalation makes provenance and authenticity systems essential to maintain cognitive security.