Using Large Language Models to help us engage the public

Many researchers in Condensed Matter Physics and Materials Science find it extremely difficult to translate their research into a form which is accessible to the public. While Large Language Models such as ChatGPT, Copilot, Gemini or Claude can provide significant help to tune your writing to engage the public, there is not a community consensus of what might be acceptable. In this talk, we will describe the various levels one might use these tools. Our typical mode of reporting on our research is to use too many buzz words, arcane language, and concepts (such as band structure or crystal symmetry) which are extremely abstract. This makes writing for the public a very challenging and sometimes time consuming task. Moreover, if you consider the rate at which students receive degrees in STEM disciplines, without tuning your writing, you may likely connect with only 5% of your potential audience.

AI tools such as ChatGPT can be used at the simplest level to quickly analyze your writing to determine what educational background is required to understand it. At the highest level of usage it can take a technical report and write a narrative which is targeted to an educational background you determine. In this talk, we will describe how these new tools might be used in a way which matches your own personal comfort level, and at the same time provide you insights in how your own undergraduate or graduate students might use these tools in the future. We will describe the various ways in which we have attempted to use ChatGPT as curators of the website FunsizePhysics whose target audience is an enthusiastic middle school student, or someone who graduated from high school. This is an important section of the population that we need to engage in an accessible way, and the use of AI tools can potentially be useful in this effort.