Designing Mindful Machines: Assuming Responsibility for the Mistakes our Machines Make
Think that you’re being objective about how you’re training your startup’s AI model? Think again. Whether it’s the data you choose, the sources you’re pulling from, or the features you’re including, all are steps where bias can be introduced. So if your data automatically supports an underlying hypothesis, you should instantly have your guard up. During this fireside chat, Sift CEO Jason Tan unpacks these complexities and outline several proactive steps startups can take to make sure bias doesn’t happen in the first place—or course-correct if and when it does. He’ll also answers questions about his path from engineer to CEO and balancing product and culture development through stages of growth.