The internet is flooded with AI-generated content, including scientific content. Unfortunately, this can reduce the quality of scientific articles, introduce inaccuracies, and weaken their ability to educate. In this #FEMSmicroBlog, Sarah Wettstadt calls for Good AI Working Habits and discusses how AI fits into science communication to create engaging and high-quality content. #CulturePlate
The issue with using AI to create science communication content
As a science communicator and writer, I am seeing the direct impact of AI tools on my work. Many blog posts (whether of scientific or non-scientific topics) read the same, with images showing flashy, dramatic, or inspiring content.
However, a second read often reveals errors, misconceptions, or missing details. Readers may be misled or receive incorrect information based on what AI deems correct or logical. Additionally, with the increasing spread of low-quality and unverified pseudo-scientific articles, people trust science even less. This has led to some people making decisions that are harmful for their health or the environment.
Nonetheless, I am convinced we can use AI tools to help people understand science in an engaging way. A study even showed that scientific publications improved with AI tools were of higher educational quality than those that received no such feedback. So, together with AI, we can indeed create better and more accessible scientific content.
We just need to be smart about it and rely on “Good Working Habits”: efficient and ethical guidelines on how to implement AI in scientific outreach.
What AI is good for in science communication
The two basic features AI tools are great for in science communication are brainstorming and proofreading. The idea is not to let the tool create content since we do not know what training data it relies on as sources.
To use it as a sounding board, ask the tool to provide critical and logical feedback. For example, when writing an article, you probably start by outlining the draft or the different sections based on your notes and experimental results.
Here, AI can support you by highlighting whether the outline is logical and whether you have covered everything from your notes. It may point out where to add more detail or restructure to improve the flow.
Once you have created a structure you are happy with, make sure to write the article yourself, using your voice and wording. AI is great at helping develop appropriate phrases, analogies, or headings for the article. Do not be afraid to ask it to explain a scientific term or concept for you; it may come up with really elegant explanations.
However, be critical when implementing its suggestions and check scientific references or your notes. Just because it sounds awesome or fits into the flow of your draft does not mean it is accurate. It may still be a hallucination.
With a first draft at hand, check with the AI tool of your choice whether you have included all relevant information from your notes and whether there are logical gaps. Ask the tool to point out jargon or tricky sections and phrases that may need improvement. Depending on the audience you want to reach and the field, consider adapting the article.
After editing your article, AI can help you with a final proofreading round. Have it check grammar, consistent spelling, and abbreviations, as well as punctuation. We all know how frustrating it is to proofread a text three times and still have one spelling mistake hidden somewhere.
But again, see the comments as critical feedback from someone reading your draft. In the end, accept only those suggestions that you deem correct and relevant.
Improving your science communication with AI
Based on an approach that sees AI tools not as writers but sounding boards, you are still producing 100% human content. Feel free to ask AI for ideas, tips, or explanations on your article, but write, edit, and adapt its suggestions according to what truly makes sense to you.
This approach allows you to produce an article that is most likely better than if you had done it on your own. External ideas and suggestions that you probably would not have come up with on your own can improve your explanations and storyline.
The end result will be a transparent, high-quality article that explains science in an engaging and relevant way. And that is exactly what each scientist (and communicator) should aim for!
Dr Sarah Wettstadt is a microbiologist-turned science writer and communicator writing for professional associations, life science organisations and researchers from the biological sciences. She runs the blog BacterialWorld to share the diverse and colourful activities of microbes and bacteria, based on which she co-published the colouring book “Coloured Bacteria from A to Z“. As science communication manager and blog post commissioner for non-for-profit organisations, Sarah writes about microbiology and environmental topics for various audiences. To help scientists improve their science communication skills, she co-founded SciComm Society, through which she offers guides, webinars and 1-on-1 coaching. Prior to her science communication career, Sarah completed a PhD at Imperial College London, UK and a postdoc at the CSIC in Granada, Spain. In her non-scicomm time, she enjoys playing beach volleyball on the sunny beaches in Spain or travelling the world.
About this blog section
The section #TheCulturePlate for the #FEMSmicroBlog aims to bring science closer to different audiences and to tell more about the scientific or personal journey to come to the results.
| Do you want to be a guest contributor? |
| The #FEMSmicroBlog welcomes external bloggers, writers and SciComm enthusiasts. Get in touch if you want to share your idea for a blog entry with us! |