Should Open Source AI Mean Exposing All Training Data?

> the existence of complete training data is [...] not an absolute requirement [...] this OSI stance is widely regarded as a reasonable approach

That's definitely NOT the case for me and it's the reason why I now see the OSI as a joke. But feel free to make up your own definitions: Windows is open-source, the PlayStation is a car, and broccolis are good for cats (for a different definition of a broccoli).