What AI Is Doing to the Way We Share Science
How we use AI to support science communication without losing the researcher’s voice or the meaning behind the message.
You may be familiar with Marshall McLuhan’s communication theory of "The medium is the message."
What he meant was that the way we share something can change the meaning of it. Take television, for example. When news moved from print to TV, it didn’t just speed things up. It made crime feel more urgent, more personal, more ever-present. That shift changed public opinion. Not because the facts changed, but because the medium did.
Now, the new medium is AI. And it’s reshaping something we care a lot about: how we communicate science.
AI isn’t just a handy writing tool. It’s an algorithm that decides how things are said, what gets repeated, and what gets left out. It tends to be confident, polished, and... often a bit samey.
That sameness is where we lose our edge. It gets harder to tell one discovery from the next, or to trust that there is a real expert behind the message.
Here’s what I’ve noticed.
1. It’s easy to lose the expert’s voice
AI sounds authoritative. But it’s not the expert. It just mimics one. If we’re not careful, we end up with content that looks right on the surface but doesn’t reflect the actual drive, passion or inspiration behind the research.
Science communication isn’t just about passing on information. It’s about helping people understand where the knowledge came from, why it matters, and what we’re still figuring out. At the centre of that story is the researcher: the person who asked the question, did the work, and is closest to the meaning of the results. The expert has a place in the story because they dedicate years to a topic. They just need a bit of help to convey their core message.
2. Everything starts to sound alike
Even when AI produces solid drafts, the tone often feels flat. When everyone uses the same tools in the same way, we lose variety. And that’s a problem. Research should feel distinct, grounded in real people and their ideas. Not templated.
Communication isn’t just about words. It’s about a sender and a receiver, each with their own beliefs, context, and filters. AI can strip all that out, if you let it.
So, how do we use AI without letting it take over the message?
Here is how our team makes sure all science communications we create feel human:
- We start with the human behind the research. We keep the scientist, researcher or STEM entrepreneur at the centre of anything we produce. We want to know who they are and why they care about their work.
- We create a persona-style profile of them, as well as the one for their audience, so everyone on the team knows who they are. Their words and intent will guide us in creating content that resonates.
- We use AI as a tool, never as a final draft. We let it research and suggest ideas that relate to the theme of the communication piece. We take it as far as we can, then move to a good old-fashioned text editor.
- Then we ask: Does this sound like a real person? Would I want to sit down and hear more from them?
- We use visuals to bring it to life. We are designers at heart and love a strong and personal look and feel. It can bring the work to life and stop it from feeling artificial. And this is something AI isn’t great at doing… yet.
- We collaborate with other creatives to spruce it up. Does it need a video, a podcast or an interactive exhibition? There is no limit to what we can do with a great team.
Science communication can absolutely benefit from AI. But only if we can get a good idea of who is speaking, and whether the message still sounds like something worth listening to. That’s why we start every project by understanding the researcher behind the science, not just what they found, but who they are and what drives them.
If you’re using AI to share research, we’d love to hear what’s working and where it’s falling short.
