Concern 795 · Working draft · editorial review in progress
Under-moderation of harmful speech
Under-moderation occurs when an automated filter fails to recognize harmful content, especially when language, context or coded meaning differs from its training examples.
Understanding the concern
A threat or hateful message may use local slang, mixed languages or an ordinary word with a special meaning in context. A system that performs well on standard English tests may miss it. At the same time, a harsh-looking quotation may be part of a report condemning abuse, so simply blocking more words is not enough.
SGHateCheck uses locally informed tests to examine moderation weaknesses in Singapore's languages. The lesson is like relying on a security guard who cannot understand much of the conversation. Local expertise, meaningful reporting routes and review of actual errors are needed alongside automated screening.
An example
A community repeatedly reports a locally understood threat that the platform's filter treats as harmless text.
Illustrative unless identified as a documented case.
Words in plain English
- Training
- The process of using examples to shape how an AI model behaves. Learning patterns from material does not establish that the material is true or fair.
Evidence and its limits
The benchmark identifies weaknesses in selected models and test cases. It does not measure all real-world harms on a platform or establish one universal moderation threshold.
WOAH 2024 · Functional benchmark
SGHateCheck: Functional Tests for Detecting Hate Speech in Low-Resource Languages of Singapore ↗ (opens in a new tab)Local-language tests refined by native annotators reveal moderation weaknesses. Not a global platform harm rate.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
A question to discuss
Which languages and forms of abuse are being missed, and can affected people obtain a useful response?
Keep track of different views, what needs checking, and any next step in your own discussion notes.
Open discussion notes →Principles for the conversation
Before you finish · reflect together
Virtues and principles to consider
Virtues are qualities we practice in how we treat people and make choices. They are not scores for judging someone. Use these ideas to explore what a thoughtful response could look like.
Unity
Unity means working toward shared purposes while making room for real differences. It does not require everyone to have the same experience, beliefs, or opinion.
Look for a goal people can share, then explore disagreements about how to reach it. Make sure quieter or less powerful voices can influence the result.
Ask together: What purpose can we share while respecting our differences?
Listening
Listening means trying to understand someone's experience before preparing an answer. People who use a service, work with it, or are judged by it may notice problems its designers miss.
Invite people to describe what happened in their own words. Repeat back what you understood and let them correct it. A quiet person or a complaint that is hard to measure may still reveal something important.
Ask together: Whose experience have we not heard yet?
Dignity
Dignity means recognizing each person's worth regardless of usefulness, status, or a system's score. People deserve respect and a voice in matters that affect them.
Explain important decisions, avoid humiliating treatment, and keep room for a person's story. Do not treat a prediction as a complete account of who someone is.
Ask together: Does this use treat each person as someone whose experience matters?
Editorial ratings
These are editorial judgments from the register, not measured probabilities or predictions about your situation.
About the register and scoring →More ways to explore
Search for more writing
For more specific or recent material, open a search below or ask for a web summary. AI-generated summaries can contain mistakes; open the original articles to check them.