Artificial intelligence algorithms often require large amounts of data to be trained, and open data can be an attractive resource to do so, due to its abundance and relative freedom to use. However, it is not always clear whether using open data for training purposes is legitimate, especially from an intellectual property rights perspective, since no permission is collected from the creators of those inputs, and AIs do not necessarily consider licensing terms.
Moreover, it is debatable what the status of the outputs from AI are: are they in the public domain, or can anyone claim ownership of the AI created work? Does it matter for ownership that the AI was guided by a human using it as a creative tool?
In this panel discussion, we discuss these and more questions related to AI generated data.