Today’s performance of artificial intelligence (AI) heavily depends on its training data, for which the donation of data by users is an important criterion. However, it is still difficult for users to anticipate how the quantity and quality of training data may affect them. Thus, users face challenges choosing between giving data to companies or keeping it confidential. That is, foregoing their privacy rights in favor of the "greater good", i.e., better AI systems not only for themselves but for everyone. In this paper, we provide a conceptual understanding paired with empirical evidence on the impact of donating data of different quality on the AI system's performance. We focus on two common data: medical data and data from entertainment. Furthermore, we discuss ethical concerns within this context. This work is not normative; rather, it empowers people to reflect on their moral beliefs and understand their impact on AI.