There is an understandable preoccupation with the technical consequences of artificial intelligence (AI) for medical writing. Large language models (LLMs) can now generate protocols, clinical study reports, summaries, literature reviews and scientific manuscripts with remarkable speed. They can restructure information, improve grammar and produce apparently sophisticated scientific prose.
But while much attention is directed towards the technical skills that medical writers must acquire to remain relevant, rather less is said about the human skills that have always distinguished a good writer from someone who simply knows how to write. This was the focus of the EMWA conference presentation I took part in this week [1].
Communication, critical thinking, emotional intelligence, negotiation, empathy, adaptability and the ability to work collaboratively are sometimes dismissed as soft skills. The description is misleading. In scientific communication, these capabilities are anything but peripheral. They are fundamental to ensuring that the right information is obtained, interpreted, challenged and ultimately communicated accurately.
Research provides an interesting perspective. Analysing approximately 70 million job transitions and skill requirements across nearly 1,000 occupations, the researchers identified a nested structure in professional skills. Advanced, specialised capabilities depend upon more fundamental skills, including communication, comprehension, critical thinking and teamwork. These foundations are not simply complementary to technical expertise; they help make its development and application possible [2].
Although this research was not specifically about medical writing, its implications are relevant. Becoming an accomplished medical writer requires much more than learning document formats, regulatory guidelines and scientific terminology. These are specialised capabilities built upon a foundation of human interaction and intellectual judgement.
Our argument as presenters was that although AI may increasingly reproduce the technical outputs of medical writing. It does not follow that it can replace the human capabilities underpinning the process.
From author to conductor
Perhaps the most important change in the medical writer's professional identity is the transition from being principally a producer of documents to becoming a conductor of the scientific communication process. A clinical study report, for example, is rarely the product of one person's understanding. It brings together investigators, clinicians, statisticians, pharmacologists, regulatory specialists, data managers and sponsor representatives. Each possesses a different perspective, vocabulary, professional interest and interpretation of the evidence. The medical writer must understand these perspectives, identify where they agree, recognise where they conflict and establish a coherent account that accurately reflects the available evidence. This is an organisational and interpersonal challenge as much as a writing exercise.
Research into teamwork identifies leadership, mutual performance monitoring, adaptability and team orientation as important components of effective collaboration. These are supported by shared mental models, communication and mutual trust [3]. The medical writer occupies an interesting position within this environment: often without formal authority over the contributors, yet responsible for helping them work towards a common document.
The analogy of a conductor is useful. A conductor does not play every instrument, nor necessarily possess the greatest technical expertise in every section of the orchestra. Their contribution is to understand the individual capabilities, establish timing and coordination, resolve inconsistencies and ensure that everyone contributes towards a coherent performance, giving the final product a clear and distinctive voice [4].
Similarly, an experienced medical writer knows when to ask a statistician to explain an unexpected result, when to challenge a clinician's interpretation, when to seek clarification from a sponsor and when apparently contradictory opinions require further scientific discussion. These activities demand confidence, diplomacy, curiosity and judgement. They are not readily automated by generating another paragraph.
The clear message from the presentations was that as routine drafting becomes increasingly automated, orchestration becomes more valuable. The challenge is no longer simply producing a document, but deciding what should be automated, what requires expert review, who needs to be involved in key decisions and how scientific quality is maintained throughout the process. The writer's contribution increasingly lies in directing the workflow rather than performing every individual task.
Listening is a scientific skill
One of the most underestimated capabilities in medical writing is listening. Not simply hearing what someone says, but understanding what they mean, what they have omitted and why they might be reluctant to express a particular opinion. A statistician may identify an inconsistency without appreciating its clinical significance. A clinician may be convinced by an interpretation that is not fully supported by the analysis. A sponsor may have a preferred narrative that needs to be examined against the evidence. The medical writer must create an environment in which these differences can be discussed constructively.
Research into psychological safety demonstrates the importance of creating working environments where individuals feel able to ask questions, acknowledge uncertainty and challenge established thinking without fear of embarrassment or punishment [5][6]. Although much of this research comes from organisational and healthcare settings, the principle translates directly into scientific document development. A writer who makes contributors feel comfortable admitting that they do not understand a result, or that an earlier interpretation may have been wrong, is likely to obtain better information than one who simply circulates a list of outstanding comments.
Effectively a humble-leadership model, it requires emotional intelligence, patience and an ability to distinguish criticism of an idea from criticism of the individual expressing it.
The medical writer does not need to be the cleverest person in the room. They need to recognise when somebody else knows more and be sufficiently confident to ask them to explain it.
The value of empathy, diplomacy and gratitude
Scientific teams are made up of people working under considerable pressure. Investigators have clinical responsibilities, statisticians have competing analytical commitments, sponsors have development milestones and writers have deadlines. A successful medical writer will navigate these competing demands without allowing the document to become a casualty of organisational friction.
Empathy helps writers understand the pressures experienced by contributors. Diplomacy allows them to pursue difficult questions without creating unnecessary conflict. Negotiation helps reconcile competing priorities, while adaptability enables them to change approach as a project develops. Even something as apparently simple as expressing gratitude at a critical point has a role.
Research that reviewed over 91 studies involving more than 18,000 participants, found a moderate positive association between gratitude and prosocial behaviour [7]. Although this does not establish that thanking a reviewer will improve a clinical study report, it provides evidence that recognising the contributions of others supports cooperative behaviour.
In practical terms, acknowledging a statistician who has gone beyond the original request, recognising an investigator's time or thanking a colleague for identifying a difficult inconsistency helps build the relationships on which collaborative work depends. These behaviours are not superficial niceties. They contribute to a working environment in which people are more willing to share knowledge, offer assistance and raise concerns.
Critical thinking, curiosity and intellectual independence
If AI can produce fluent scientific prose, the ability to question what that prose actually means becomes increasingly valuable. A medical writer must be prepared to ask whether the available evidence supports a conclusion, whether an apparent association has been interpreted correctly, whether an alternative explanation exists and whether an important limitation has been overlooked. This is not simply quality control. It is scientific reasoning.
Studies have clearly demonstrated how convincingly ChatGPT could generate apparently plausible scientific abstracts, including abstracts containing entirely fabricated research data [8]. The distinction between a well-written scientific statement and a scientifically defensible one is therefore fundamental.
The philosophy of science offers another useful perspective. Work on scientific objectivity emphasises the importance of critical interaction between individuals with different perspectives. Scientific knowledge becomes more robust through opportunities for assumptions to be challenged and evidence to be examined collectively [9]. The medical writer can facilitate precisely this process. Rather than merely recording consensus, they can help expose unresolved questions and encourage contributors to examine the reasoning behind their conclusions. This requires intellectual independence and, occasionally, the courage to be considered inconvenient by the rest of the team who want to ‘move on.’
It also requires curiosity. A good medical writer does not accept that a result is interesting simply because a senior member of the team says it is. They want to understand why it matters, what it means and what remains uncertain.
This is why scientific depth matters more than ever. Medical writers who understand disease biology, study design, endpoints, statistics, regulatory expectations and clinical practice contribute something beyond document production. They provide context, interpretation and scientific judgement. As AI becomes more capable of generating text, expertise becomes increasingly important because it allows writers to recognise when a conclusion is unsupported, a claim is overstated or an important limitation has been overlooked.
AI makes the human contribution more visible
There is a danger that the increasing availability of AI tools will encourage organisations to define medical writing principally in terms of document production. If the objective is simply to generate 100 pages of reasonably coherent text, automation will appear attractive. But the document is only the visible product of a much larger intellectual and collaborative process. AI can assist with summarisation, drafting, restructuring and editorial refinement. It can also help identify inconsistencies and suggest questions for further investigation. Used appropriately, these capabilities can release writers from some repetitive activities and provide more time for scientific engagement.
However, the writer must still determine whether the output is accurate, appropriate, complete and consistent with the source evidence. Current professional recommendations also retain human responsibility for the accuracy and integrity of scientific work, irrespective of the use of AI [10].
The challenge is to avoid becoming an editor of machine-generated text who has lost contact with the reasoning behind it. As discussed in Niche's earlier commentary on AI-assisted writing, the apparent solution to writer's block can simply relocate the problem from generating text to verifying and owning its meaning [11][12][13]. Writers need to understand the many limitations associated with LLMs. Niche prepared a special edition Insider’s Insight for the meeting, describing the current landscape (though be aware this will date quickly) [14].
The future medical writer must therefore combine technical AI literacy with the confidence to question its output and the interpersonal skills to bring human expertise into the process. Increasingly, organisations are not paying medical writers simply to produce documents. They are paying for ownership. Someone must define the objective, evaluate the evidence, identify risks, challenge assumptions and remain accountable for the final output. AI can contribute to the drafting process, but responsibility for the scientific integrity of the work remains human. The writer's value lies not in generating words, but in standing behind them.
Developing the human side of medical writing
Soft skills should not be treated as personality traits that someone either possesses or lacks. They can be developed through practice, feedback and reflection. Research into deliberate practice highlights the importance of targeted activities, repeated performance and constructive feedback in developing expertise [15]. For medical writers, this means deliberately practising more than document production. Seek opportunities to lead scientific discussions. Ask questions during investigator meetings. Learn to explain difficult concepts to people outside your immediate discipline. Practise giving and receiving constructive criticism. Develop the confidence to challenge assumptions respectfully.
Another increasingly important capability is process fluency. Medical writers need to understand how scientific information moves from protocol to report, submission, publication and communication. This includes recognising where AI can accelerate routine activities, where it introduces risk and where human judgement must remain firmly in control. Writers who understand the entire process are better positioned to direct it effectively and identify weaknesses before they become problems.
One key point emerged during the course of the discissions; how will the medical writer of tomorrow get the training they need in an environment that lessens the burden on your executive skills [16]? Perhaps medical writers should also spend more time writing outside their routine professional assignments. Research into the distinctive voices of health researchers highlights the importance of developing a repertoire that extends beyond conventional academic structures [4]. Writing commentaries, exploring scientific controversies and expressing independent opinions can strengthen intellectual confidence and communication. So is the feedback you get from your peers. Mistakes are often the best training experiences.
These activities also help writers develop something increasingly valuable: an identifiable professional perspective.
The future belongs to communicators
It was generally agreed that temergence of AI does not make technical expertise, regulatory knowledge or writing ability redundant. They remain essential. But they may no longer be sufficient to distinguish an exceptional medical writer from an increasingly capable automated alternative. The greater opportunity lies in recognising that medical writing is fundamentally a human activity concerned with communicating scientific knowledge.
The most valuable writers will be those who can connect people, interpret perspectives, challenge assumptions, build trust and guide complex discussions towards a scientifically defensible outcome. They will combine scientific understanding with critical thinking, process fluency and professional ownership. They will be recognised not simply for producing documents, but for helping organisations make better decisions.
To stay relevant, you need to become facilitators, negotiators, critical thinkers AND communicators. You will understand when to use AI, when to question it and when human judgement must take precedence.
Perhaps the irony is that, as machines become better at writing, we may finally recognise that the greatest contribution of a medical writer was never simply the ability to put words on a page, that was just the bit your colleagues struggled with. It was the ability to understand the science, challenge the reasoning, navigate the people involved, take responsibility for difficult decisions and ensure that complex evidence was communicated honestly and effectively. In an age of AI, those capabilities are the human writer's greatest competitive advantage.
References
- Lepage R, Singh N, Hardman TC. How to AI-proof your medical writing career. EMWA 62ndConference, Autumn 2026.
- Hosseinioun M, et al. Skill dependencies uncover nested human capital. Nat Hum Behav. 2025;9(4):673-687.
- Salas E, Sims DE, Burke CS. Is there a "Big Five" in teamwork? Small Group Res. 2005;36(5):555-599.
- Lingard L, Watling C. The writer's voice repertoire: Exploring how health researchers accomplish a distinctive 'voice' in their writing. Med Educ. 2024;58(5):523-534.
- Edmondson AC. Psychological safety and learning behavior in work teams. Adm Sci Q. 1999;44(2):350-383.
- LaPlante RD, et al. Essential elements and outcomes of psychological safety in the healthcare practice setting: A systematic review. Appl Nurs Res. 2025;83:151946.
- Ma LK, Tunney RJ, Ferguson E. Does gratitude enhance prosociality?: A meta-analytic review. Psychol Bull. 2017;143(6):601-635.
- Gao CA, et al. Comparing scientific abstracts generated by ChatGPT to real abstracts with detectors and blinded human reviewers. NPJ Digit Med. 2023;6(1):75.
- Longino HE.Science as Social Knowledge: Values and Objectivity in Scientific Inquiry. Princeton (NJ): Princeton University Press; 1990.
- International Committee of Medical Journal Editors. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. Updated January 2026.
- Hardman TC. (2026) Medical Writing 2026: Adapting to AI and Rising Complexity.
- Hardman TC. (2026) What Every Good Medical Writer Knows.
- Hardman TC. (2025) Medical Writing vs Artificial Intelligence: Threat, Tool, or False Debate?
- Niche Science & Technology Ltd. (2026). What Differences in LLMs Mean for Scientific Authors.
- Ericsson KA. Deliberate practice and acquisition of expert performance: A general overview. Acad Emerg Med. 2008;15(11):988-994.
- Hardman TC. (2026) Who Trains Tomorrow’s Scientists?
