Human-Centered AI: Lessons from Building Two LLM-Powered Products

Explore key insights from building two LLM-powered products in this 25-minute conference talk by Margaret Jennings from Kindo. Learn about the rapid evolution of large language models like ChatGPT and their transition from research to mainstream products. Discover how AI systems enable "active design" through user feedback and how to quickly demonstrate value by leveraging LLM strengths such as summarization and retrieval. Understand the importance of designing for human-AI collaboration by simplifying complexity and making LLMs more accessible. Gain knowledge on how LLMs condense the product design process, using each interaction to inform future iterations. Examine the need for thoughtful co-design of models and interfaces to shape the societal impact of LLMs. Consider the potential of open, inclusive design and proactive policy in steering LLMs to empower humanity. Benefit from key lessons learned to guide the thoughtful development of LLM-powered products in the future.

Human-Centered AI: What I've Learned From Building Two LLM-Powered Products Thi... Margaret Jennings