Artificial Intelligence Set for Big Role in Scientific & Technical Publishing

Artificial Intelligence Set for Big Role in Scientific & Technical Publishing

Intelligent technology one day becoming ubiquitous, leaving us all jobless and at the mercy of machines, is a notion deeply rooted in science fiction for most people. But truth is sometimes stranger than science fiction; much of intelligent technology has already been developed, and scientific and technical publishers are some of the current leaders in deploying it.

In its latest report, Global Scientific & Technical Publishing 2017-2021 Simba Information examines how STM publishers use AI to solve practical problems, develop tools for deep learning, and help navigate industry disruption tied to digital transformations and new open access business models. 

Although these technologies show great potential for scientific and technical publishers and researchers, the debate about how much intelligent technology is too much is only beginning. 

The McKinsey Global Institute (MGI) made headlines in January when it released an eye-opening report on automation technologies and their potential effects.

“Our scenarios suggest that half of today’s work activities could be automated by 2055, but this could happen up to 20 years earlier or later depending on various factors, in addition to other economic conditions,” according to MGI.

The report found that the activities most susceptible to automation are physical ones in highly structured and predictable environments, as well as data collection and processing. In the U.S., these activities make up 51% of activities in the economy, accounting for almost $2.7 trillion in wages.

Current advances in automation technology mean robots and AI programs can perform more tasks that humans typically perform, including data analysis and collection.

More Scientific Articles Being Published 

Simba estimates that there are more than 2.5 million scientific articles published every year in English, the dominate language for science, and the rate of growth is close to 5%.

But that is just the published papers. The glut of research that STM publishers must evaluate is much larger. Research spending grows at 3% annually, meaning for every published article, countless more are outright rejected or kicked back to the authors for refinement.

Many publishers view artificial intelligence as a means of bringing more speed and efficiency to operations. STM publishers already use some early AI technologies to augment and automate tasks involved in peer review, including:

  • Identifying new peer reviewers with broader searches
  • Fighting plagiarism with software that can identify components of whole sentences or paragraphs rather than verbatim text
  • Reporting when authors fail to report key information needed to make the decision to accept or reject a paper
  • Finding bad statistics that can lead to false conclusions
  • Ferreting out data fabrication.

Establishing AI's role in determining which research to publish is where things get tricky for publishers. 

As publishers introduce more AI-powered checks into the review process, the need for a human review of artificial intelligence's results grows. AI can misinterpret data sets; it can raise red flags on items for which there is a logical explanation. Without some sort of human review, potential mistakes from artificial intelligence mistakes could prevent publication of insightful scientific literature.

STM publishers and the research community will deploy AI to combat the consequences of the industry's move to more open publishing models: predatory journals. Most research funders now mandate that articles produced as a result of their grants be made available for free.The path to open access is paved by upfront fees called Article Processing Charges (APC) paid by the author or funding body to make the article freely available on the Internet. Reputable publishers are deploying this model, but so are many fly-by-night operators who look to take an author’s money and publish their work without critical review.

AI solutions will address the rise of predatory journals in two ways. AI will increase the credible scientific and technical publishing community’s capacity to publish quality works by helping editors find new reviewers, and by creating automated reviews. AI technologies will also make it possible to automate reviews of published works to identify journals and publishers that are not fulfilling their obligation to uphold standards.