TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Twitter's image-crop AI can favor white men, women's chests

3 pointsby aarestadover 4 years ago

1 comment

rbeckerover 4 years ago
&gt; This was based on a few publicly cited examples of Twitter&#x27;s photo framing using images of both light-skinned and dark-skinned people that skewed toward the light-skinned person. [..] And Zehan Wang, engineering lead at Twitter<p>For all their talk about the importance of diversity, why did the complaining Twitter users only check White-Black bias, and not Asian-Black as well? Especially since the engineering lead is Asian, and Twitter&#x27;s diversity report claims Asians are over-represented (27% vs US 5.4%) and whites under-represented (42% vs US 61%) at their company: <a href="https:&#x2F;&#x2F;blog.twitter.com&#x2F;en_us&#x2F;topics&#x2F;company&#x2F;2019&#x2F;Board-Update-Inclusion-Diversity-Report-May2019.html" rel="nofollow">https:&#x2F;&#x2F;blog.twitter.com&#x2F;en_us&#x2F;topics&#x2F;company&#x2F;2019&#x2F;Board-Upd...</a><p>&gt; Vinay Prabhu, chief scientist at UnifyID and a Carnegie Mellon PhD, ran a cropping bias test on a set of 92 images of white and black faces and found a 40:52 white-to-black ratio, which argues against bias for that particular set.<p>So how many test images did the people that found bias use? Because this could be just cherry-picking, as only the cases where bias was found get attention.