Human-Centered AI Narratives: How a 12-Million Image Dataset of Skin Shades is Evolving AI's Ubiquitous Microcosm
カートのアイテムが多すぎます
カートに追加できませんでした。
ウィッシュリストに追加できませんでした。
ほしい物リストの削除に失敗しました。
ポッドキャストのフォローに失敗しました
ポッドキャストのフォロー解除に失敗しました
Human-Centered AI Narratives: How a 12-Million Image Dataset of Skin Shades is Evolving AI's Ubiquitous Microcosm
-
ナレーター：
-
著者：
このコンテンツについて
AI's rapid ubiquity is exciting and threatening, especially to those outside of its development. Emmanuel Acheampong, the visionary co-founder of yShade.ai, has meticulously curated a dataset of 12 million diverse skin tones. This feat isn't just transforming the beauty industry's approach to inclusion—it's poised to enhance safety standards for autonomous industrial machines, prevent fatal accidents in self-driving vehicles, and revolutionize dermatological diagnostics.
With host Jessica Buchleitner, Emmanuel challenges the prevalent Value Neutrality Thesis of technology—the widely accepted narrative that technology emerges as neutral and becomes imbued with traits based on its user behavior. Through compelling examples, he reveals AI as a ubiquitous microcosm, where small enclaves of engineers and technologists create solutions for global populations while primarily addressing scenarios relevant to their limited perspectives.
What sets his approach apart is his human-centered methodology of building a diverse AI dataset through qualitative research interviews with beauty industry experts and medical professionals. This provides a stark contrast to the narrative that AI datasets are built solely by scraping the internet. As DEI measures disappear in government and private companies, he explains how economic forces remain the beauty industry's driver for inclusivity.
One conclusion is apparent from this conversation: Our understanding is fundamentally limited by our datasets, and we don't know what our data fails to show us.
Connect:
- Emmanuel's LinkedIn
- yShade.ai