In this episode, we delve into a multi-layered approach to evaluating the risks of generative AI, as outlined by DeepMind. From misinformation to systemic bias, we explore the tools we need to make AI safe in real-world applications, and how gaps in current evaluations can lead to unforeseen societal harm.

Download: https://deepmind.google/discover/blog/evaluating-social-and-ethical-risks-from-generative-ai/