Righting writing’s wrongs: Toward effective writing partnerships between humans and AI

In industrialised nations, we take for granted that knowledge is acquired and shared through a written medium, and yet the actual task of writing remains a hard skill to master. In this paper, we present a blueprint for the integration of artificial intelligence (AI) into Web-based text editors, and show that this contributes towards simplifying the writing task for authors. Our approach utilises the textual content of a document as input into existing artificial intelligence processes, extended with an RDF knowledge graph. The supplementary material from the AI partner is presented to an author for possible inclusion in their document. We separately evaluated two approaches to the presentation of linguistic tone analysis into editors. Our results from unmoderated observational user tests validate the premise that cognitive technology can simplify the writing task, and show that design is critical to the adoption of cognitive technologies. We conclude with lessons learned from deploying informal qualitative research sessions and online observational tests, and share opportunities for future work. We note that merely applying AI or any new technology to a human activity is patently insufficient. Our limited experience and evaluation suggest that design and user testing are necessary activities for adoption and acceptance