Dr. Mirko Bischofberger is an expert in science communication. He is a lecturer at the Swiss Federal Technology Institute of Zurich and Lausanne (ETH and EPFL), as well as the universities of Zürich and Basel. Mirko studied molecular biology, specialised in computational biology and worked for the Neue Zürcher Zeitung, a renowned Swiss newspaper. He will speak at the upcoming ESMH Summer School ‘Science Communication in the Digital Age’ held on June 24-26, 2025.
What first sparked your interest in artificial intelligence (AI) and science communication?
Mirko Bischofberger: I think it comes from the fact that I studied bioinformatics as part of my studies, so I was always a bit of a geek. And so in the last few years, AI has been growing, and I got interested in the field, and started to use my skills to try to get a feel for what AI really is. And then ChatGPT arrived, and I was fascinated. So, the short answer is that it comes from my interest in computer science.
What are the opportunities of using AI to communicate science?
Mirko Bischofberger: I think the most interesting thing happening at the moment is the field of AI persuasion. Recent studies have shown that AI, especially large language models, can make people more likely to change their beliefs. AI systems can interact better with humans to make them see different versions of an issue. Maybe AI systems are better listeners, making the person feel more trusted and welcomed, and therefore also more trusting of new information? Or maybe AI systems have simply the better arguments.
AI offers a promise of a better way of communicating. The capacity to help reach beyond the usual audiences is one of the greatest opportunities of AI for science communication, I think. It could help give access to scientific knowledge to people who are usually not accessing it. Concretely, if you have somebody who believes the world is flat and you put them together with a person or with a machine, the machine is more likely to change their mind. It is only just happening, but it says a lot about the potential of AI and communication.
Persuasion could potentially also pose significant risks if AI gains the capacity to manipulate people. What do you see as the greatest risks with using AI for science communication?
Mirko Bischofberger: Of course, dual technologies like AI can be used for both good and ill. It comes with great advantages, but also great risks, including from deepfakes and misinformation. Managing these risks will be key.
It’s the same with using AI for science communication. Science communication has always been at the border of a big tension in our field, which is about simplifying science but not misrepresenting it. With AI, this problem will be exacerbated. Because machines will hallucinate, lie, and maybe persuade people of falsities even better.
Can we trust LLMs, considering their propensity to hallucinate?
Mirko Bischofberger: It’s definitely a problem. But I have seen a lot of error-prone communication: humans are not perfect either. And once AI technologies get to a certain level where, statistically, they’re better than humans at the given task, they become good tools. What we are seeing now is the worst version of science communication with AI that history has ever seen. It can only get better.
AI can also help communicate a broad range of scientific fields, setting it apart from humans. Most communicators hardly know one field well. So I think we have to have a certain humility and ask: what do we really know about science? AI tools can help us, but there will always be errors.
In your teaching on science communication, you emphasise the importance of storytelling. How can stories help communicate science?
Mirko Bischofberger: It’s a beautiful paradox. On the one hand, AI technology is new, and social media has been new 20 years ago. On the other hand, some things stay the same. Since the dawn of humanity, storytelling has been used as a way of packaging information. Because humans love stories. They love having a hero, a protagonist with a challenge. So I think there is still potential in using storytelling better to make a message more compelling and memorable, to help people identify with a scientist and their journey. Storytelling has always been around, but we still are not using it enough in our field.
Is there a risk that storytelling misrepresents the nature of scientific endeavor?
Mirko Bischofberger: Definitely. It comes back to this balance of simplification and not portraying something too wrong. I mean, explaining science in simple terms already has this problem. So adding a layer of a story makes it even more difficult. So yes, there is a risk that you put the story too much at the center. But there are also opportunities, which is that people can finally access and identify with science.
What role do you see for things like images and film in communicating science?
Mirko Bischofberger: It depends on the message. First, you have to choose what you want to tell, what’s the science, what’s the story, what’s the level of language in terms of the audience, and then you choose the format. Films and videos have had a huge growth in the last few years due to social media. But I sometimes see too many videos. So it’s not always the best form of media. Sometimes it’s still good to have something simple, like a blog entry or an infographic. Just because everybody can do videos, but that doesn’t mean that you always need a video.
How can events such as the ESMH Summer School help promote better science communication in Europe?
Mirko Bischofberger: Every summer school, which is an idea born in academia, brings people together. You network and get to know other people who are interested in the same thing. You bring people up to a certain level. But ideally, we don’t just learn and meet but also think about the future and come up with new ideas. Participants can go in new directions, have new ideas, and dive into possible futures of science communication.