The Role of AI and Human Expertise in the Future of Research Communication
The publishing of scientific research has been transformed by the introduction of artificial intelligence (AI) into nearly all aspects of the research workflow. Authors and researchers at all career stages use AI tools for proofreading, paraphrasing, citation management, literature discovery, and text summarization. But beyond this, how is AI contributing to the communication of research? And where does human expertise matter? Let’s dive in!
Process of Scientific Publishing
Role of AI in Research Communication
Future of AI and Human Experts in Scientific Research Communication
Process of Scientific Publishing
Here’s a simplified flowchart of various stages in scientific publishing.
Manuscript preparation: In this stage, authors gather research data, perform experiments and analyses, and report their findings in a manuscript that is written and formatted as per the target journal guidelines.
Initial submission: The authors submit the paper to the journal for publication.
Editorial screening: The editor-in-chief or the academic editor of the journal screens the submitted manuscript to see whether it aligns with the journal’s aims and scope. The manuscript is also checked for compliance with journal guidelines and may be desk rejected if it does not satisfy the requirements.
Peer review: The manuscript is assigned to expert reviewers in the field, who evaluate the originality, novelty, quality, and significance of the research. The peer reviewers provide feedback and recommend acceptance, revision, or rejection.
Revision: The authors rework on their manuscript based on the reviewers’ suggestions and editor’s feedback. The revised manuscripts are resubmitted to the journal along with suitable responses to reviewers’ comments.
Acceptance: The manuscript is accepted for publication if it meets the journal’s standards. In this stage, the journal will work on typesetting, copyediting, and proofreading the manuscript and the authors may need to approve the final versions before publication.
Publication: The research is published in the journal, either in print, online, or both, and made available to the readers.
AI tools play a significant role in several stages of this entire workflow. But research is no longer about how it is presented; it is also about how it is communicated.
Role of AI in Research Communication
Once your research is published in a journal, there are many ways for you to promote it on various platforms. From creating graphical abstracts to obtaining plain language summaries, AI is known to simplify the process of research communication.
- Visual communication: AI-powered tools like BioRender and Mind the Graph are popular choices for creating scientifically accurate graphical abstract that help in visual communication of scientific research. Authors need not be designing experts or professionals to create eye-catching graphical abstracts. With a vast selection of graphics and illustrations that cater to various scientific fields, these tools provide authors design templates that can ease the process of creating graphical abstracts.
- Plain language summaries: Jargon-free plain language summaries help policymakers, patients, or general audiences understand the implications of scientific research. And AI-driven writing aids like HypeWrite AI, Paperpal, and Grammarly can be useful assistants for authors in this process.
Moreover, readers can get summarized text or bullet points highlighting the key aspects of a lengthy research paper to consume research content easily. - AI for translation: Sharing knowledge with a global audience is no longer a challenge because of AI translation platforms. This benefits both authors and readers, considering that the former gain recognition for their work and the latter can easily access scientific findings without having to worry about language barriers.
- Simplified literature search: Tools like R Discovery, LitMaps, and Research Rabbit utilize AI to keep track of published scientific research. When scholars, students, and researchers use these tools, features like audio papers, text translation, and collaboration of reading materials makes it extremely convenient for them to access your research.
Nevertheless, no AI tool can replace the critical thinking abilities of humans, and their oversight remains crucial when handling these tools.
Future of AI and Human Experts in Scientific Research Communication
A large part of the scholarly community emanate skepticism when it comes to using AI for scientific research, and they are not entirely wrong. But the problem is not the use of AI—it’s the overuse of AI and relying on AI output without questioning its quality and integrity.
- Who can forget the AI-generated illustration of a rat with disproportionate genitalia1 and misrepresented text? This mishap could have been easily avoided had the people responsible taken the time to review the graphics rather than blindly rely on AI tools!
- More recently, a strange phrase occurring in research papers2 was traced back to a mistranslation of a term in a paper published decades ago. “Vegetative electron microscopy,” which is clearly a nonsensical terminology, went unnoticed by researchers, authors, and peer reviewers for years, leading to misguided scientific research. A deep dive into this blunder revealed that a bad digital scan of two papers published in the 1950s mistakenly combined “vegetative” and “electron” from two different columns, resulting in this erroneous term. Since then, it has been inaccurately translated to different languages by GPT models, highlighting the risks of unchecked AI in scientific research.
- There are even instances of AI tools generating falsified information, such as authorship or citations3, misleading researchers with references to works that did not exist. This is because the tools are trained to always give you an answer, and when they fail to find the information you are looking for, they tend to fabricate data.
Do you still think that human intervention is unimportant? Think again!
There are many such instances of AI mistakes that have underscored its limitations. Which is why experts in the industry always recommend cautious usage of tools with adequate human oversight. And if you are wondering where human expertise should intervene when using AI tools, the answer is simple: Everywhere!
- The developers working on AI tools need to ensure that the information provided to large language models (LLMs) is accurate. Moreover, the configuration and training of AI tools must be carefully monitored to avoid generation of false, non-existent information.
- Companies and organizations managing AI-powered platforms should emphasize updating the software on a timely basis and keep eliminating any bugs in the backend.
- Users must educate themselves or get training from the right sources before using AI tools for scientific research. For instance, crafting clear, concise, and effective prompts is a skill that will come in handy in the future.
- Scientific researchers must do what they do best: Question the integrity of AI-generated information. Always verify authenticity of the source of the information before presenting it in a research paper.
Key Takeaway
Yes, AI tools are great assistants when integrated into the research workflow, but that’s all they are—assistants. And that’s why the future of scientific research communication will need to be a balanced integration of human + AI.
References
1. Study featuring AI-generated giant rat penis retracted https://www.vice.com/en/article/ai-midjourney-rat-penis-study-retracted-frontiers/
2. A strange phrase keeps turning up in scientific papers, but why https://www.sciencealert.com/a-strange-phrase-keeps-turning-up-in-scientific-papers-but-why
3. Formatting with AI may be riskier than you realize https://www.cliffedekkerhofmeyr.com/en/news/publications/2025/Practice/Knowledge-Management/knowledge-management-alert-20-june-formatting-with-ai-may-be-riskier-than-you-realise
Break that writer’s block! Check out Paperpal and simplify your academic writing with AI assistance.







