The conversation around Plain Language Summaries (PLS) has become increasingly focused on artificial intelligence.
Can AI generate an initial draft? Can it reduce authoring effort or accelerate timelines? These are reasonable questions. But after spending time with sponsors, transparency teams, patient engagement specialists, patient advocates, and regulatory stakeholders, I believe the industry may be focusing on the wrong challenge.
| PLS is not fundamentally a writing problem. It is a transparency workflow problem.
Until we address it that way, we risk optimizing one step in the process while leaving the larger challenges untouched.
The market has become fixated on draft generation
Many of today’s PLS discussions begin with a Clinical Study Report and end with a generated document. The workflow typically looks something like this:
| Source document > AI > PLS draft
The assumption is that if we can generate a draft faster, we have solved the problem. But drafting was never the most difficult part of PLS. The real challenge begins after the first draft is produced: clinical review, quality review, patient and lay review, translation, approvals, traceability, version control, consistency across studies, and governance.
Generating content is only one activity within a much larger ecosystem. Yet much of the industry’s current innovation effort remains focused on that single step.
Patients do not consume information the way we think they do
Perhaps the most overlooked aspect of the PLS discussion is how patients actually consume information. The industry often talks about simplifying language. Patients care about understanding information. Those are not necessarily the same thing.
A four-page summary can still be difficult to absorb. A well-written paragraph can still be less effective than a simple graphic. A technical explanation can become far clearer when supported by a visual timeline, diagram, iconography, or interactive explanation.
Yet many PLS solutions remain overwhelmingly focused on text generation. The implicit assumption is that the final output is a document. That is too narrow a view. The real objective should be comprehension: how quickly patients can understand a study, how easily they can navigate the information, how confidently they can interpret the findings, and how accessible the overall experience is.
Patient review should be more than a PDF and a comment sheet
This becomes even more important when we consider patient engagement. Regulators encourage meaningful patient involvement in patient-facing communications, yet many review processes remain surprisingly static. A document is distributed. Comments are collected. Feedback is consolidated. Versions are exchanged. The process repeats.
But why should patient engagement be limited to reviewing a PDF?
Patients and patient advocates provide some of the most valuable perspectives in the process. They identify terminology that is unclear, highlight assumptions that researchers may overlook, and help determine whether explanations are genuinely understandable to the intended audience.
Their contribution should not be treated as a final validation step. It should be part of the workflow itself.
Imagine a review experience where patient advocates can provide structured feedback directly within the content, comment on graphics and visual explanations, identify areas of confusion, participate earlier in development, and collaborate alongside other stakeholders rather than after the fact. That is a fundamentally different approach to patient engagement, and one that technology should make easier, not harder.
The scalability challenge nobody talks about
Even if AI successfully accelerates drafting, another question remains: can the process scale?
Most organizations are not creating a single PLS. They are creating many across studies, therapeutic areas, regions, languages, and review teams. The challenge quickly stops being about writing and becomes one of consistency, governance, review management, and institutional knowledge.
If every PLS begins as a completely independent effort, organizations inevitably recreate work that has already been done. Terminology is debated again. Review cycles are repeated. Visual assets are recreated. Stakeholder feedback is rediscovered. Decisions lack continuity.
The problem is not authoring. The problem is fragmentation, the same challenge transparency teams have been managing for years.
Why PLS should not live in a separate world
What is striking about the current market is how often PLS is treated as a standalone process. By the time an organization begins creating a Plain Language Summary, significant work has already taken place. Clinical source documents exist. Disclosure planning is underway. Confidential information decisions may already have been made. Participant privacy and anonymization strategies may already have been established. Governance processes are already in motion.
So why does PLS so often start from scratch? Why are these activities managed in separate systems, workflows, spreadsheets, review cycles, and repositories of knowledge?
The future should not be another disconnected process. It should be a connected model where source documents, disclosure, confidential information management, participant privacy, anonymization, patient communication, patient review, governance, and AI-assisted content creation share a common foundation.
Not only because it is more efficient, but because it creates consistency, traceability, scalability, and trust.
AI is important. But it is not the story.
AI absolutely has a role to play. It can accelerate drafting, identify relevant source content, support consistency, and reduce repetitive effort. But the future of PLS will not be determined by who generates the first draft fastest.
The organizations that lead will be those that create trusted, scalable, and patient-centered processes around that draft, bringing together technology, governance, patient engagement, human expertise, and transparency operations.
| Patients do not benefit from AI-generated content. Patients benefit from understanding.
And understanding requires much more than a language model.
The bigger opportunity
The more I think about the future of PLS, the more I believe we have been asking the wrong question.
The industry keeps asking, “How do we generate Plain Language Summaries faster?”
A better question is: “How do we create a transparency operating model that helps patients understand clinical research more effectively?”
Those are very different challenges. One produces documents. The other builds trust. And in the years ahead, trust may prove to be the most important outcome of all.
