Ethics
The speakers in CQC chose to speak publicly. Collecting their voices into a labeled research corpus still creates new risks, so we limit what we record, what we share, and how it can be used.
Principles
- Based on self-description. We annotate only what speakers said about themselves.
- Not stated does not mean absent. Missing information stays missing; we do not fill it in.
- Identity is not a sound. CQC documents variation; it is not a dataset for classifying people.
- Public is not unrestricted. A video being public does not make it appropriate to redistribute it freely in a labeled dataset. Annotated data and audio go only to approved researchers.
- Speakers can request data removal. Material is removed on request, and approved users must delete it too.
Safeguards
- Before each release, every source is checked again. Material whose video has been deleted or made private is removed.
- Every user signs the Data Use Agreement, which prohibits identity inference, re-identification, contacting speakers, and voice cloning.
- We keep a record of every approved user so that removal notices reach everyone who holds the data.
The project was reviewed by the PI’s institutional review board. In designing CQC we took into account China’s Personal Information Protection Law and copyright law, as well as the laws that apply where the corpus is held.
Request removal
If you believe your voice is in CQC and you would like it removed, email us. You do not need to give a reason.
| Stage | What we do |
|---|---|
| 1. Received | We respond to your email as soon as practicable, in Chinese or English. |
| 2. Withdrawn | We stop distributing the material immediately and remove it from the next release. |
| 3. Users notified | We notify every approved user, who must delete it within 30 days under the agreement. |
What removal covers: the material is taken out of every future CQC release, and approved users must delete their copies. We cannot withdraw group-level results that researchers have already published, because these do not contain the material itself.
Write to chinesequeercorpus@outlook.com. A link to your channel or video is enough for us to find the material.
Email a removal request中文说明请见 移除请求。