Certain English Female Contributor Submitting False Data

Hey there,

I started contributing to common-voice today and noticed a certain english female contributor which has been submitting a lot of false data.

She usually either says only part of the provided text or says something completely different.

It’s now gotten to the point where I hear a clip of her about every 4-6 times. It’s gotten somewhat infuriating.

Is there any way an admin of some sort could prevent her from contributing?

there is a report button at lower left corner of validation interface, if you can report some of them probably someone can batch invalid the other.

AFAIK, the best way to prevent spammers like this will be invalidating the recordings (by two persons). The whole workflow is based on this.

The REPORT button currently only saves the sentences & sentence ID’s to reported.tsv, it is not related to the voice, it is about the text-corpus. Even so, this info is not taken into account in the current workflow (i.e. in CorpusCreator, Coqui etc).


Some other ways to handle this would be:

  • Giving language communities more rights/responsibilities
  • Speaker based statistics - e.g. if a person’s recordings got invalidated more than a given percent, he/she will be blocked (temporarily/permanently)
  • Downvote button for speakers (risky)

These do not exist (yet)…

1 Like

Alright. I suppose i’ll just “no” / “downvote” the recordings in question instead of reporting them then. Thanks!