Why we need it?

Hi everyone!

Thanks for giving full information for my last inquiries. Now I have a vision of how to start the project or in other words, challenge. I read the instructions and almost finished localizing Common Voice web page. We, our NGO, do not see it as a problem to unite volunteers to scale up and complete the mission.

However, we have come across some questions:

  1. what we benefit from Common Voice (no financial, of course)?
  2. what if we finish the project with minimum results?
  3. where and how we can use these databases? ( e.g Google/Yandex Voice typing)

Terribly sorry, if I rewrote questions asked before.


Personally, I’m contributing because I pre-ordered a mycroft.ai and this project is the best way to speak to it without also speaking directly to an NSA collection facility. Currently, all the good speech to text options are run by large American corporations that don’t care about privacy. Fixing this is a way to take some power back.

1 Like

If you use voice systems from Google, Amazon, Apple etc, your voice data will be used to train the algorithm whether you like it or not. This may include personal or sensitive information about you. While Amazon instructs employees to flag up personal data so it won’t be used, a human still has to listen to it to determine that it’s personal in the first place.

With Common Voice you know exactly what data you are contributing and sentences are random from an encyclopedia. It allows you to improve voice algorithms with a whole lot more consent and control, without sacrificing a lot of personal data.

Additionally, large companies will only invest in languages that make sense for them commercially. Common Voice is probably the only way that smaller languages actually have a chance of being supported by voice technology.

1 Like