Unmasking AI: My Mission to Protect What Is Human in a World of Machines Audiobook by Joy Buolamwini

Developing the Aspire Mirror

• 💡 Joy Buolamwini was developing an Aspire mirror, a project involving a video filter that appeared on a mirror reflection.
• 🚀 Her initial proof of concept was successful, but she aimed to add interactivity by making the image follow her face as she moved.

Encountering Technical Hurdles

• 🛠️ To achieve interactivity, she needed to integrate face tracking software with a webcam.
• 🧩 Like many coders, she relied on pre-existing software libraries rather than building everything from scratch, comparing them to prefabricated building blocks.
• ⚠️ She integrated an open-source face tracking library into her code for the project.

Unmasking Algorithmic Bias

• 🔍 Despite looking directly into the camera, the face tracking system could not detect her face.
• 🎨 To test the system, she drew a simple face on her hand, which the software successfully detected.
• 🎭 Crucially, when she wore a white costume mask, the software detected her "mask face," but when she removed it, her dark-skinned human face was not detected.
• ⚡ This personal experience revealed a significant algorithmic bias in the face tracking technology.