Small Is Sexy: Rethinking Article Length in the Age of AI

Abstract

With the emergence of large language models (LLMs) like ChatGPT, the scholarly community faces a growing question: Are lengthy articles still the best way to communicate research? Recently, Pividori (2024) highlighted the ability of AI tools to generate large volumes of text quickly, prompting reflection on whether long articles truly advance scientific progress. As we move further into the AI era, should we instead focus on brevity and clarity?

This article is an opinion piece that reflects on the evolving challenges in academic publishing, particularly in the context of the increasing role of AI tools like LLMs. We aim to provoke thought and inspire action towards adopting concise and impactful scientific reporting in response to the growing issue of information overload in the research community. This piece deliberately reflects on the ethical, environmental, and academic implications of publishing in the AI era. It does not claim to present empirical findings but rather serves as a reaction to current developments and an argument for rethinking traditional publishing practices in favour of clarity, efficiency, and sustainability.

The current scientific information landscape is overwhelming. Jinha (2010) estimated that since the 17th century, over 50 million scientific papers have been published, with more than 2.5 million new papers added annually. This flood of content makes it nearly impossible for researchers to stay fully informed, even within their fields. Lengthy papers, often filled with excessive detail, contribute to this overload, demanding more time and cognitive resources from readers.

In today's fast-paced research environment, efficiency and clarity are critical. Adopting a ‘small is sexy’ approach, inspired by Hill's TED talk ‘Less Stuff, More Happiness’ (Hill 2011), could improve the quality of academic writing. Reducing verbosity allows more focused, impactful research findings to shine without being buried under unnecessary information. This shift can lead to better comprehension and faster dissemination of critical insights.

Some may counter our argument, claiming that detailed methodologies and large datasets are essential for reproducibility and understanding. While this is true, balancing necessary detail with conciseness is possible. One solution is to include Supporting Information or external repositories for exhaustive datasets and methodological specifics. By placing such information in accessible repositories like Zenodo or Dryad, researchers can ensure the integrity of their work without overloading the main text.

This approach has multiple benefits. It simplifies reading, enhances transparency, and promotes open science. Researchers can focus on interpreting results and discussing their significance rather than getting bogged down in exhaustive descriptions. It also facilitates interdisciplinary collaboration, making research more accessible to those outside the immediate field who may not need every technical detail but are interested in broader findings.

For publishers, supporting repositories not only helps reduce the length of articles but also improves the reproducibility and impact of research. By integrating supplementary datasets into publications, journals can offer a more streamlined reading experience while ensuring that all necessary information is available for those who require it.

The ethical implications of using AI in scientific writing further support the case for brevity. Moffatt and Hall (2024) express concerns about AI's inability to take responsibility for intellectual contributions. If AI can generate long, seemingly impressive texts without truly understanding the content, relying on such tools to inflate article length undermines the integrity of scientific communication. Instead, AI should assist in processing and synthesising information, helping researchers distil key insights rather than producing lengthy, unfocused documents.

Furthermore, long articles have environmental costs. Producing, distributing, and digitally storing large texts consumes energy and resources. According to The Shift Project (2019), digital technologies account for about 4% of global carbon emissions, which is expected to double by 2025. Reducing the length of articles can help decrease data storage demands and energy consumption, aligning scientific publishing with global sustainability goals.

By embracing concise reports, publishers can contribute to more eco-friendly practices, reducing the carbon footprint of both physical and digital distribution. This aligns with the growing push across industries to meet sustainability targets while improving the efficiency of information sharing.

The ‘publish or perish’ culture in academia often values quantity over quality, pushing researchers to produce more work, sometimes at the expense of originality. The ease with which AI can generate text may worsen this issue, leading to an influx of lengthy but shallow publications. This overloads the peer review system, as reviewers must sift through additional material that may offer little new insight.

By shifting towards concise, impactful reports, we can ease the burden on reviewers and improve the quality of academic contributions. Shorter, more focused papers allow reviewers to concentrate on the essential aspects of research, leading to faster turnaround times and more meaningful feedback. In turn, this can improve the quality of the published literature.

Embracing this shift could also lead to more efficient editorial processes for journals. Fewer, shorter submissions mean less strain on resources and faster publication times, which benefits authors, reviewers, and readers alike.

Ultimately, this shift towards brevity forces us to reconsider the purpose of scientific writing. Is the goal to impress with volume and complexity or to communicate ideas clearly and effectively? In an era of information overload and limited attention, clarity and conciseness are more valuable than ever. By distilling research to its most impactful elements, we increase its accessibility, influence, and ability to inspire.

For scholarly publishers, prioritising quality over quantity means promoting research that contributes meaningfully to the field rather than simply adding to the growing mountain of information.

The rise of AI language models presents an opportunity to rethink how we approach scientific publishing. Long articles may no longer meet the needs of the research community, contributing to information overload, wasted resources, and environmental harm. By adopting a more concise approach, utilising data repositories for Supporting Information, and embracing short reports, we can improve the clarity, efficiency, and sustainability of academic publishing.

In a world overloaded with information, embracing brevity can make scientific literature more meaningful and accessible, benefiting both researchers and society.

The authors contributed equally to the conceptual design, literature review, writing, and editing.

The authors declare no conflicts of interest.