037 – Generating AI Images: Between Bias and Absurdity

In this follow-up to episode 035, Lily and David consider a recent development in the realm of AI image generation: Google’s Gemini model recently came under criticism for generating historically inaccurate images, seemingly as a result of overcorrecting for racial biases. They discuss topics including: how making mistakes can show that Google might be ahead of the competition; new AI regulations from the EU; Elon Musk’s lawsuit against OpenAI; and the infamous Glasgow Willy Wonka experience. Read about the Google Gemini incident here: https://www.theverge.com/2024/2/21/24079371/google-ai-gemini-generative-inaccurate-historical