Research on the Construction of Information Disclosure Mechanism for AI-Generated Content
Generative artificial intelligence poses significant challenges to the governance of information content.The current information disclosure model for AI-generated content,which centers around prominent labeling,has limited capacity to carry information.It over-looks the information disclosure obligations of entities other than service providers,which is detrimental to the secure and controllable development of generative AI products.Based on considerations of the purposes and target audiences for information disclosure at various stages of'model training-content generation-information dissemination',the elements of information disclosure for AI-generated content include technical information,rights description information,AI contribution information,and warning information.In terms of the de-sign of information disclosure procedures and accompanying measures,AI-generated content information disclosure should establish query-based and quasi-rights management information disclosure models according to different scenario requirements.This should clarify the responsibilities for inaccuracies in user disclosures and the proactive identification obligations of dissemination platforms.By promo-ting technical interoperability among entities,the full lifecycle management objectives for AI-generated content disclosure information can be achieved.
万方数据
维普
