Artificial intelligence (machine learning, deep learning, neural networks, and algorithms) has the potential to free-up humans to be more creative, to explore new research areas and to change education provision. A case is in publishing, where humans and technology are co-existing and are ‘working in partnership’.
Artificial intelligence can aid publishers with:
- Automated text analysis: Technology can detect plagiarized content without human input and thus reduce editorial workload.
- Content translations: Creating professional translation automation tools, removing the risk of human error.
- Content personalization: For transitioning to customized content curation within newsletters and audience segmentation.
- Automated formatting: Process like editing, formatting, and checking grammar can be automated by algorithms.
- Search engine optimization: This includes actions like automating keyword research, performance optimization, and distribution.
- Text auto-tagging: Enabling the use of metadata tagging to improve the discoverability of content.
As an example, Writefull, which produces artificial intelligence based language services, has been integrated into the American Chemical Society’s publications workflow. This is through
Writefull’s application programming interfaces becoming integrated for three key uses.
The API is used before copyediting to automatically classify all accepted manuscripts by their language quality. The API assigns a level-of-edit grade to manuscripts at scale without editors having to open documents and review the text.
The American Chemical Society is an important scientific publisher, producing more than 300,000 research manuscripts in more than 60 scholarly journals each year.
The artificial intelligence technology is trained on millions of scientific papers using Deep Learning. Through this, the algorithm identifies potential language issues with written texts, offers solutions to those issues, and automatically assesses texts’ language quality. The process can be applied at all key points in the editorial workflows.
To test out the algorithm, this involved benchmarking alongside human editors. In these tests, Writefull reached more than 95 percent alignment in grading texts.
Writefull’s Metadata API can also be deployed to automate aspects of manuscript review, ensuring that all elements of an article are complete prior to publication.
The API is also used by the open access publisher Hindawi as a pre-submission structural checks tool for authors. In this context the technology checks whether all authors have affiliations, key sections are present, and all references are cited in the text.
