Artificial intelligence (AI) powered text generation will fundamentally change scientific publishing. In recent years, multiple AI systems have showcased the production of visual and textual content that is increasingly indistinguishable from human-generated work, creating almost overnight new possibilities for intellectual workers and, at the same time, raising similarly potent concerns. While artists and journalists are more evidently at the forefront of this incipient revolution, it is not hard to imagine a researcher looking away from the frustratingly sparse draft of a research article and wondering: "Could a machine write it for me?"

This question might have passed for a flight of fancy until recently, as machine-generated scientific arguments were easily distinguishable from human output, and paper-generating software mainly highlighted the permeability of the peer-review process from nonsensical process to nonsensical papers. However, these technologies have progressed so rapidly that we have likely entered a new phase in which machine-generated text can be seamlessly integrated into human-generated scientific articles. 

AI and scientific publishing

Several experts who track down problems in studies believe that AI's rise has turbocharged the problems in the multi-billion-dollar sector. All the experts emphasized that AI programs such as ChatGPT can help write or translate papers if they are thoroughly checked and disclosed.

It is not always easy to spot the use of AI. However, one clue is that ChatGPT tends to favour certain words. Andrew Gray, a librarian at University College London, trawled through millions of papers searching for the overuse of words such as meticulous, intricate, or commendable. He determined that at least 60,000 papers involved the use of AI in 2023—over one per cent of the annual total.

Meanwhile, according to the US-based group Retraction Watch, more than 13,000 papers were retracted last year, the most in history.

Revolutionizing boundaries

Recent breakthroughs in the use of AI in complex strategic games have highlighted the surprising ease with which AI can outcompete humans in problems considered intractable with computational approaches. Similar breakthroughs in the use of AI for scientific progress might come from the combination of (i) precise goals, i.e., a clear definition of what we consider a successful scientific observation, (ii) algorithms capable of efficiently optimizing its output for these goals, and (iii) structured and accessible scientific data. 

In this line of thought, we can imagine AI systems proposing new experiments and descriptions of observed phenomena and arranging data in figures to support their conclusions. An AI system capable of producing original scientific work could revolutionize the whole scientific endeavour, for example, by being less tied than humans to the boundaries of scientific disciplines, bringing multidisciplinary science to new heights.

Phys Org