MovieSum's card says CC BY-NC 4.0 but not where the screenplays came from, and the HF API record has no license tag. SummScreen states no license and draws on TV transcript sites. kahuja/flawed-fictions is tagged MIT with an empty README. A statement from the dataset authors or the original text rights holders, or a provenance section in the papers' full text, would show whether the labels cover the underlying texts.
Why this is still open: Only abstracts, cards and READMEs were read; paper bodies and the rights holders' terms were not.