Post from your feed
“This comment was framed as a joke, but it was targeted, gendered, and meant to make me leave the conversation.”
public reply
possible harm
Relaunching with new features on 25 May
See what the machine sees. Then correct it.
VoiceValor shows people how big tech classifies their social media feed, lets them correct what the classifier gets wrong, and invites them into a community that decides how this data should be used.
Why it matters
Misogyny, harassment, and context-specific abuse often look invisible to automated systems. VoiceValor makes those gaps visible and correctable.
Social feed
Classifier view
@your-feed
Today
Another post
“The platform called it low severity. People in my community knew exactly what it meant.”
coded language
context missing
Big tech classification
The model marks the post as low-risk because it misses local context.
Low severityYour correction
Choose what the system should have seen.
Correction sent back to the classifier.
Community layer
You can join others to decide who gets access to the corrected data and what support is needed.
How it works
VoiceValor makes systematic change possible
See the feed through big tech’s eyes
People can view their social media feed as a classifier sees it: the labels, risk scores, categories, and blind spots that shape whether harm is noticed or ignored.
Correct what the classifier gets wrong
When a label is wrong, incomplete, or too shallow, people can correct it. Those corrections become feedback that can be sent back to improve the classifier.
Join a community that governs the data
People can become part of a community that decides who should get access to the data, under what conditions, and how contributors can be supported too.
Impact so far
What changes when people can correct the system.
It is possible: people closest to online harm can reveal what automated moderation misses, especially when abuse depends on language, culture, gender, caste, religion, conflict, or political context.
0
survivors and marginalized internet users contributed lived evidence.
0
platform detection failures identified in culturally specific online misogyny.
0
data donations gathered through survivor-led, consent-based participation.
0
women and girls in South Asia reached through safer digital experiences.
From lived harm to evidence
“VoiceValor has been one of the few spaces where we can talk about conflict, about hate, about violence, and more importantly, do something about it.”
Shared by a survivor from Sri Lanka who contributed to VoiceValor’s evidence-building work.
Safer platforms begin with the people they failed first.
VoiceValor is part of Dignity in Difference’s broader work on digital safety, feminist technology, and platform accountability across South Asia.
Visit Dignity in Difference