The AI Readout: How LLMs report the latest clinical science and why this matters

The AI Readout: How LLMs report the latest clinical science and why it matters

Click the image to read the full report

What happens to new clinical data once it enters the world of AI?

We wanted to find out. 

To better understand how new scientific information is surfaced through generative AI, we tracked how four major large language models (LLMs) reported a new clinical trial data announcement during the month following its presentation at a European medical congress.

Using a range of prompt types, we examined when the data first appeared in AI-generated responses, which sources the models appeared to rely on, and how those responses evolved over time.

The results point to a new reality for healthcare communications

Our research identified three important trends:

  • New data can surface quickly.
  • Owned content can be highly influential.
  • The wider media ecosystem can shape whether AI considers data relevant.

However, the story does not stop at publication day.

Our research suggest that visibility evolves over time, meaning the way organisations communicate new science after the initial announcement may be just as important as the announcement itself.

In the report, we’ll unpack three findings: 

  • Why recency matters in AI visibility 
  • Why your owned website still matters 
  • Why the breadth of media coverage can influence what AI surfaces 

Because in an AI-driven health ecosystem, publishing the science is only the beginning. 

Want to understand what this could mean for your next data moment? Get in touch with Richard Buchanan Brown or Hannah Richards.

Download the full report here