Harnessing AI to strengthen trustworthy information
summary
Demonstrates a Lucene‑based A2 search stack and an integrated RAG assistant that returns sentence‑level citations, plus experiments with Wayback Machine, vector image search and zero‑shot classifiers. Three curated search engines (WHO/CDC, top clinics, and influencer blogs) answer the same queries and produce strikingly different outputs — celery juice as harmless hydration versus miracle cure, vaccines flagged as safe by official sources but uncertain in the blog set — and the demo shows the RAG summaries citing sources and working as intended.
AI-written summary · about
chapters (8)
- 0:00Introduction and speakers
- 1:52Challenge of misinformation
- 6:12Reliable sources tools
- 12:57Demo on credible sources
- 17:37How answers get generated
- 23:55Expanding search and context
- 27:05Classifying information
- 34:18Key takeaways and Q&A
related talks
21mEvaluating AI Search: A Practical Framework for Augmented AI Systems
19mBuilding a Smarter AI Agent with Neural RAG
43mAdvanced Retrieval-Augmented Generation Techniques
51mTransforming Banking with Advanced AI: GenAI, RAG, and Multi-Agent Systems
44mRAG for a Medical Company: Technical and Product Challenges AI & Data
01h 48mInformation Retrieval from the Ground Up
26mAccelerating COVID-19 insights for clinical use with intelligent search
19mImproving RAG Systems - Key Challenges and Solutions