Under the NINJAL collaborative research project, we started the compilation of a large-scale corpus of everyday Japanese conversation, the Corpus of Everyday Japanese Conversation (CEJC). The CEJC targets various kinds of naturally occurring conversations in daily situations, such as conversations during eating with the family at home, meetings with colleagues at work, and conversations while driving. A notable characteristic of the CEJC is that both audio and video data are collected and published. However, as virtually no corpora have contained video recordings of everyday conversations, guidelines have to be established on the release of such data in light of the potential ethical and legal issues of showing the faces of non-consenting third parties or copyrighted content such as TV programs. Based on a variety of data collected thus far, we discussed with a lawyer specializing in copyright and portrait right issues how to deal with such ethical problems. In this paper, we report a guideline for the release of the CEJC that we have established based on this discussion.
『日本語日常会話コーパス』データ公開方針 : 法的・倫理的な観点からの検討を踏まえて
papers1505.pdf
(1.13MB)
概要
