Translation of creative content and technology

Translation of creative content and technology

This panel focused on the intersection of autonomous human creation and machine-assisted creation of content, which may be used or (partially) de-selected when delivering documents and products across different media and institutional settings in the creative industries. 

Different contexts may serve as exemplars. For example, neural machine translation (NMT) trained on literary data does not appear to have the necessary capabilities for creative translation; it renders literal solutions to translation problems and, more importantly, using NMT to post-edit raw output constrains the creativity of translators, resulting in a poorer translation (Guerberof-Arenas & Toral 2022). In audiovisual translation, machine translation post-editing has become a common practice, without always serving as a guarantee of the intrinsic creative component required for translators. Media-propelled controversy on successful international series (e.g. Squid Game 2021, Hwang Dong-hyuk) regarding subtitle output in streaming platforms points at an increasing awareness of the translation act and the reception of translation. In other sectors of the creative industries with both a wealth-generation and a community focus, such as museums, texts and narratives are made available through different media and technologies, often entailing moments of mindful mediation for their respective audiences. 

Overall, there is a sense that machine translation models, tools and media need to be designed or used with translators (and their audiences) as a central focus. In this way, default, routine output may be backgrounded and technology can foster creativity rather than constrain it. 


Dr Dimitris Asimakoulas

Dimitris Asimakoulas

University of Surrey