Unexplainable Good Robot #2: Everything is not awesome
137 snips
Mar 15, 2025 
In this discussion, AI research scientist Margaret Mitchell, linguistics professor Emily M. Bender, and Vox's Seagal Samuel examine the unsettling consequences of AI. They delve into ethics and responsibility when AI causes harm, reflecting on real-world impacts on minority populations. The trio also tackles biases in technology, the challenges of emotional complexity in AI, and the urgent need for ethical oversight. Their candid insights stress the pressing concerns over AI's future and highlight the balance between innovation and societal responsibility.
AI Snips
Chapters
Books
Transcript
Episode notes
Everything Is Awesome
- Margaret Mitchell trained an AI on Flickr images, leading to an "everything is awesome" problem.
- The AI labeled a factory explosion "awesome" due to the vibrant sunset colors in the images.
Kismet Inspiration
- Joy Bulamwini's interest in AI began as a child watching a PBS show about Kismet.
- Kismet, a social robot at MIT, captivated her with its expressive eyes and emotive abilities.
Coding in Whiteface
- Joy Bulamwini, while coding a digital face mask, noticed facial recognition AI struggled to detect her dark skin.
- The AI readily detected a white mask, revealing a bias in the technology.
