Jodie Burchell's keynote, "Can you trust your (large language) model?", from NDC AI 2025, tackles a crucial question: how reliable are those awesome machine learning models we're increasingly using? Even though they crush complex tasks, figuring out if their outputs are truly correct isn't as simple as it seems.
Especially as "black box" LLMs become commonplace, understanding when and how these models might be leading us astray is super important and increasingly tough. This talk promises to dive into how our models can actually "lie" to us, offering ways to cut through the confusion and get to the real truth.
Watch on YouTube
Top comments (0)