Safely Managing Generative AI with Humans in the Loop

Explore two real-world case studies demonstrating human-in-the-loop (HITL) approaches to safely managing generative AI in this 30-minute conference talk from the Toronto Machine Learning Series (TMLS). Learn about BlueDot's global disease surveillance team using large language models (LLMs) to identify concerning infectious disease outbreaks, and Moov.AI's development of question-and-answer services incorporating a knowledge base. Discover the challenges faced in ensuring factual, specific, and appropriate AI-generated content while managing reputational risks. Gain insights into the design patterns involving human oversight in production-ready generative AI systems, aligning with regulators' push for human oversight of "high risk" systems. Understand the implications for creating strategic advantages and the lessons learned from these use cases for organizations considering generative AI applications.