Guiding Students to Use AI to Build Science Writing Skills

As the world of technology continues to evolve and advance, it becomes increasingly important to enhance our students’ exposure to opportunities that will build their ability to use technology safely and effectively. One way to do this is by leveraging generative artificial intelligence (AI) in a project.

As a science teacher who loves to write, I wanted to bring writing projects into my classroom. However, my students did not share my enthusiasm, and I wanted to give up until I tried a new approach with AI. All my students had heard how AI was basically like cheating, and although I acknowledge that some students have abused this technology, the power of this new tool is exciting and motivating, in my educational opinion. So, I decided to introduce an AI writing project as a science experience: Students would learn how to use AI effectively and with fidelity while also practicing their writing and research skills.

Introducing an AI Science Project

When it was time to introduce the AI project to my students, I began by explaining that this would be a new type of creative assignment that would combine our science content with the technology they all had been talking about—AI—but I would show them how to use this technology in an ethical manner. First, I went over the expectations I had created that had broken each part of the project down into manageable chunks. The basic parts of the project were as follows:

• A typed, double-spaced research paper with multiple sources
• An AI component
• A three-to-five-minute oral presentation

Next, I stunned them by doing something I had never done before—I informed them that I, too, would be completing the assignment. They were excited, and shocked, that I would be doing this project with them.

After the shock of that announcement wore off, I dove into the specifics of the project: I wanted students to select a person from science whom most people had never heard about before. Although the scientists in our textbooks had accomplished noteworthy things, there are other individuals who deserve to be recognized too, so I wanted my students to focus their attention on those not included in our textbooks.

As educators, it is important to expose our students to the reality that there are many more scientists, and specifically many more who can better represent the diverse students we teach. I chose to research Kathryn D. Sullivan, NASA’s first woman to do a spacewalk, because I am the biggest NASA nerd ever.

When I finished my soapbox speech on that topic, I challenged my students to take the rest of our class time to get on their computers or go to the stack of biographical books I had checked out before class to broaden their knowledge of diverse scientists and help them choose which scientist to research.

After introducing the project, I also taught them how to use AI platforms such as ChatGPT, Grok, and Microsoft Copilot to ask questions like the following:

• List scientists who have made a contribution who are Mexican American.
• List scientists who have made a contribution who are Native American.
• List scientists who have made a contribution who are African American.
• List scientists who have made a contribution who are Chinese American.
• List scientists who have made a contribution who emigrated from Germany.
• List scientists who have made a contribution who emigrated from Italy.
• List scientists who have made a contribution who emigrated from India.

Because my class is very diverse, I wanted to allow my students the opportunity to select a scientist who represented their culture if they wanted to.

Empowering Students to Get Creative

The next day, students had to submit a research proposal contract stating who their scientist was, why they chose the person they did, and what questions they would ask them if they could. Finally, they signed a statement at the bottom of the page that acknowledged they would use a combination of the internet, AI, encyclopedias, and books from a library to locate information ethically. This reinforced the notion that all writers need multiple sources and viewpoints when composing any piece of writing to avoid bias, especially when using AI.

Once the contract was signed, it was their time to get to work. I turned the students loose to start conducting research, and to my delight, they were actually focused and excited about the work they were doing. By empowering students to choose their own topic, I created a classroom full of eager researchers. They spent the next few days conducting research using various sources and asking deep and philosophical questions:

• What made you choose science as a profession?
• What achievements are you most proud of in your career?
• How does your work help others?
• What were your toughest obstacles, and how did you overcome them?
• Who were your mentors in the science community?
• How do you give back to the next generation of scientists?

I knew this project was more than just another assignment when they wanted to add an extra component to the project. This group of students was pretty tech-savvy, and they had the idea to create AI avatars to help present their project. I didn’t even know what that would look like, but as I sat there speechless and stunned, all I could do was nod yes. To be honest, I was scared because I had no idea when I would find the time to do this part of the project ahead of my students, but I allowed myself to be vulnerable and told my class my concern.

One of my techy kids said he would teach me—talk about a full-circle moment. Luckily, we have Canva, and their avatar demos make it easy to create lifelike images of the person you are presenting. While all students had to turn in their research papers to me, they used Canva to allow their avatar to answer questions about their scientific life.

On the day of the presentations, science came alive in our room. Students were asking AI avatars like Sally Ride and Mary Jackson what it was like to be women in NASA’s space program and the challenges they faced, and AI was producing real-time, accurate simulated data that made one feel like these two pioneers were in the room with us.