Overview

DALL-E 2, OpenAI’s AI engine that generate images from simple text descriptions was released earlier this year. With its amazing capabilities, also came a lot of questions. How do we release this safely? How do we filter the model’s outputs without overstepping? How do we deal with social biases depicted in images? What’s our role in protecting artists whose work can now be duplicated? Learn from a panel of product, ethicists, and researchers that worked with the first Microsoft products that shipped with DALL-E 2 – Bing and Designer, to discuss the challenges and learnings along the way in integrating generative models into our products. This session will be hosted by Mark Van Hollebeke with Jeff DeVries, Nathan Freier, Forough Poursabzi, and James Zhou