You have full control. Data generation is opt-in, on a per-project basis, and you can pause data generation whenever you want.
The data is generated by plugins called DataSources, which are community-made and open-source. Since the community owns the data generation code, there is less of an incentive to make something untrustworthy, and it's a lot more difficult since the code is open source.
Transparency is key here. I think that once we have a transparent and fair foundation for sharing user-generated data, we can build some really cool stuff with data that was never possible before.
This sounds like it could open huge possibilities in a lot of industries, provided there's transparency.