With the rapid adoption of artificial intelligence (AI) across publication workflows, there is a need for a unified approach to how, where, and why we implement AI within the publication development process.
AI adoption in publications has outpaced shared standards for its use. This is a problem. Without clear guardrails, there is a risk of creating inconsistency in how AI is applied and disclosed. Varying AI use in the publication workflow is already shaping how clients assess agency credibility and trust. The publications process is long-standing and cannot be put at risk through a lack of alignment on how AI is implemented into workflows. Ethical and transparent reporting of clinical data within publications is a baseline expectation, and inconsistent or poorly defined AI use may jeopardize confidence in us as medical communications professionals.
If AI is here to stay, what guardrails must come with it?
AI has clear potential to support technical- and efficiency-related tasks for medical communications professionals, and guidance from the International Society of Medical Publication Professionals (ISMPP), Good Publications Practice (GPP) standards, and industry and congresses are emerging and evolving. However, efficiency without consistent approach is not innovating the publications workflow; instead, it is risk disguised as progress. There must be some flexibility in how AI is used as a tool in the publications workflow to align with client and author expectations, but AI efficiencies must be human-led and cannot be implemented at the expense of confidence in what we produce.
There is also a longer-term consideration. An overreliance on AI risks creating a generation of writers who can edit outputs but lack the foundational understanding to challenge them. As an industry, we have a responsibility to ensure that early-career writers have the opportunity to develop the critical skills needed to interrogate data, challenge assumptions, and uphold authorship standards, to not rely on AI-generated suggestions and outputs but to be able to use AI as a human-led tool coupled with that foundational experience.
What skills must we protect to safeguard the future of publications?
As agencies sitting at the intersection of sponsors, authors, and publishers, we are well positioned to lead by example. In our view, AI has a place in publication development, but its use must be explicit and agreed from the outset and transparently disclosed. This means embedding AI governance into existing publications processes following client and author consultation, rather than treating AI as an efficiency shortcut layered onto well-established workflows. Decisions around AI use should be intentional, documented, and reviewed in line with evolving tools and expectations.
We harness AI not to replace expertise, but to elevate it. Our approach is human-led and guided by a deliberate and transparent approach to AI in publications development. We believe in clarity around when and how AI may be used, early alignment with clients and authors, and continued focus on human accountability, authorship integrity, and scientific rigor remaining central at every stage of the process. We encourage agencies to walk this tightrope of innovation and AI stewardship, integrating AI where it adds value, while protecting the development of foundational publication skills and judgement that underpin ethical publications.
In this way, publication ethics in the era of AI are preserved and progress becomes the forefront. Grounded in intention and transparency, AI can serve as a catalyst for innovation while reinforcing the trust at the heart of our publication process.