Simple Strategies to Use AI in Undergraduate Science Education

In Blog by Rozalind Jester

By Tom D’Elia, Ph.D., Professor of Biology

Indian River State College

Advances in the sciences are evolving faster than ever, and with artificial intelligence (AI) reshaping science education, instructors are racing to keep up. Educators must revise how they support the development of the next generation of scientists, preparing them to use AI responsibly and effectively, and not as a replacement for understanding. Developing core scientific thinking, evaluating information, and building communication skills remain essential. Without structured exposure to these tools during their undergraduate years, students risk being underprepared for graduate programs and outcompeted for careers in an increasingly AI-integrated workforce.  This post outlines practical ways to integrate AI into undergraduate research experiences while reinforcing critical thinking and student-generated content.  These tips and practices were developed and implemented in a year-long undergraduate research course in biology; through the integration of AI assistance in this course, students enhanced their general understanding and expanded their skills in both data analysis and communication.   

Using AI tools for background research 

Initial literature searches can be challenging.  Students new to a topic do not know where to start.  AI tools can help students perform literature searches and process the findings into summarized key points. The resulting AI output is often interactive, with students being able to ask follow-up questions, which is one way AI can adapt to a student’s level of understanding. By using these tools thoughtfully, students can build a stronger foundation for their projects and enhance their overall learning experience. Many tools are available, and the best strategy is to explore what works for different applications and to be creative with their use.  Here are some examples that have been useful in supporting students in the classroom:

  • ChatGPT and Deep Research Agent: ChatGPT is the obvious AI tool for most students (which is available without a subscription).  This model  can summarize scientific articles and explain difficult concepts, reducing feelings of anxiety or of being overwhelmed. Students can also interact with the content by using follow-up questions to clarify areas of confusion or ask for additional examples. Interacting with the content can allow students to dig deeper as their understanding increases. The ChatGPT Deep Research agent provides a more comprehensive “deep dive” into their research topic and presents a clearly summarized overview. As students further explore their topic, they can design multi-step commands to fine-tune the reports generated by Deep Research, a unique feature of this tool. In addition to ChatGPT, other comparable models such as Anthropic’s Claude and Microsoft’s Copilot can be used for background research.
  • Google NotebookLM allows students to create personalized notebooks by uploading sources, such as PDFs, textbook excerpts, or web page links, specifically related to their projects, — then interact with that material. Features include custom summaries, quizzes, and mind maps. A popular feature among students is generating a podcast summary that they can listen to and even ask questions that are answered in real time. These features reinforce essential ideas through repetition and engagement. Many students are initially unfamiliar with this tool, but once introduced to its features, they quickly see its value and enjoy using it. 
  • Paperpal offers writing support but is also valuable during background research, helping students find relevant sources that align with their working hypotheses or project questions directly.   

Common Pitfalls: Some AI tools “hallucinate” by generating false or misleading information (Ji et al.). Use this as a teaching opportunity: have students verify sources and cross-check AI-generated content to reinforce good research habits. 

Tips for responsible use: Design assessments that ensure students have reviewed and understood the information they have collected.  Having students write summaries from the sources, identify concept connections between sources, and create lists of questions has been a helpful way to get students to review and think critically about the information they collect using AI.  Guide students to interact with the tools when they need further clarification.  Having students discuss what they are learning in small groups and with faculty also helps to check for gaps in understanding.   

Key Takeaway: AI tools can help students complete background research by simplifying complex information and guiding them to relevant sources. When used properly, these tools support critical thinking and build confidence. 

AI tools help students form their hypotheses and complete experimentation 

As students become more familiar with their research topics, they can use AI tools to brainstorm hypotheses and refine experimental designs. While lab protocols often provide a starting point, specific projects usually need adjustments.  AI chatbots can help students explore variables like replicates, reagent concentrations, or planning general experimental setup. These tools can support troubleshooting, but students still must understand the steps and determine if their plans work with available resources. Frequent check-ins with faculty help keep students on track.   

Useful AI tools for data analysis 

AI tools can help students from all skill levels to explore analysis programs that may be too complex otherwise.  Often, analysis software packages require complicated command-line installation and execution.  With AI support, students can run and interpret results from these programs more easily, even without coding experience. 

One of the most revolutionary AI applications is AlphaFold. The developers of this program received the Nobel Prize, highlighting the significance of this protein structure prediction program (Callaway). Accessible via Google Colab Notebooks, students can input amino acid sequences and receive a high-quality 3D protein model without coding requirements. With assistance from AI chatbots for simple Python commands, students can further explore protein structure, function, and evolution with tools like PyMol. Another recently developed program, CRISPR-GTP, further highlights AI’s educational potential. This tool automates CRISPR gene-editing experiments, making advanced biotechnology accessible to beginners (Qu et al.).  Exposure to these tools can spark student interest in bioinformatics and molecular biology, while providing a practical entry point into coding. 

Students can also be introduced to data analysis platforms like the open-source web-based Jupyter Notebooks, which support multiple coding languages and are widely used in undergraduate science education (Craig et al.). These tools allow students the option to edit code directly, helping them build coding skills with flexibility to match varied experience levels (Bascuñana et al.). This creates a valuable opportunity to integrate AI chat tools to guide students through coding commands and help them interpret the results of their analyses. Similar approaches using AI support are also helpful as students explore data analysis using Python, R, Mathematica, and even advanced features in Excel. 

Common Pitfalls: With so many AI tools and analysis options, students can easily lose focus or get lost down dead ends of trial and error. Faculty can help by providing guided protocols and checking in regularly to keep students on track.  

Tips for responsible use: Students must be aware that even though these programs can provide data analysis results, the output can be inaccurate if the values were input incorrectly or the wrong parameters were selected.   

Key Takeaway: Advancements in AI now enable students to use data analysis tools without needing advanced command line skills.  Since this field is growing so quickly, students may find new AI tools that can be used for their projects.  Faculty should keep an open mind to exploring data analysis programs and consider how AI can help students process their data.  This is an excellent way to update the course curriculum. 

Preparing written reports with AI assistance 

AI’s use for writing has become one of the most controversial aspects of AI in education. Students began using ChatGPT to help with written assignments as soon as it became available. While the generated text can seem impressive, it may be inaccurate or fabricated. A recent Nature Reviews Bioengineering editorial emphasizes the importance of human-generated writing in developing scientific thinking (“Writing Is Thinking” 431), while still understanding the value of AI as a tool to improve writing. As AI-assisted writing becomes more ubiquitous( from autocomplete to grammar suggestions integrated into everyday platforms), the goal should be to guide students in using it to become more thoughtful, effective communicators and scientists. 

In my undergraduate research course, students use large language model (LLM)–based AI tools at different points in preparing reports, posters, and oral presentations. For example, students are encouraged to begin writing their reports using ChatGPT to review their prepared outlines and use the feedback to organize their ideas better and logically map out their results and discussion. As students complete each report section, we have students peer review each other’s work and then use AI tools to help implement the critiques, providing suggestions for phrasing and clarity. When guiding students in using ChatGPT, we stress its responsible use, such as clarifying complex concepts, reworking awkward sentences, or formatting citations rather than generating entire text sections. Students also apply these tools when organizing presentations. In addition to supporting the initial outline of a talk or the layout of a poster, AI tools can also help students condense talking points and relate their findings to current research. Combining AI tools, instructor guidance, and peer review has created a supportive structure allowing students to build confidence in communicating scientific information.

Other AI writing tools are widely available, including Grammarly, which helps polish grammar and tone, and Paperpal, which can suggest academic phrasing, identify areas where arguments need strengthening, and help find citations to add support.  Each tool has strengths, and combining them allows students to evaluate their writing better and improve iteratively. 

Common Pitfalls: AI-generated writing can easily misrepresent a student’s writing abilities. Faculty need to set clear expectations of what is acceptable use for a given project.  They should emphasize the importance of student-generated content to minimize the development of over-reliance on AI-generated text. 

Tips for Responsible Use: Set aside time during class or lab for students to write. In-class activities that promote student-generated writing include preparing rough drafts, refining AI-generated outlines, or summarizing literature. Another helpful approach is requiring weekly updates where students present progress on their writing or presentations to peers and faculty.  These sessions often reveal whether students truly understand their work and can help spot overreliance on AI. It is important to instill in students the importance of generating their own work while maximizing the use of all tools to refine their abilities.   

Key Takeaway: Encouraging students to write regularly and review their work in group settings builds accountability and reinforces the importance of authentic scientific communication.  

A final note: 

Now is the perfect time to show students practical ways to learn content, data analysis, and writing skills without feeling overwhelmed or falling into the trap of irresponsible use of AI. Students and faculty must continue to adapt faster than ever before as AI tools rapidly get integrated into teaching and learning. Thoughtful integration of AI into classrooms and undergraduate research will enrich the learning experience and better prepare students for careers in a rapidly evolving world. 

AI Use Disclosure and Process Statement:

AI tools, including Grammarly and ChatGPT, were used during the drafting of this article to assist with grammar, phrasing, and proofreading. All content was written, reviewed, and edited by the author to ensure accuracy and originality.

 References:

  • Bascuñana, Juan, Sergio León, María González-Miquel, Enrique J. González, and Javier Ramírez. “Impact of Jupyter Notebook as a Tool to Enhance the Learning Process in Chemical Engineering Modules.” Education for Chemical Engineers, vol. 44, 2023, pp. 155–163. https://doi.org/10.1016/j.ece.2023.06.001.
  • Callaway, Ewen. “Developers of AlphaFold Win Chemistry Nobel.” Nature, vol. 634, 2024, pp. 525–526. https://doi.org/10.1038/d41586-024-03214-7.
  • Craig, Paul A., Jason A. Nash, and Thomas D. Crawford. “Python Scripting for Biochemistry and Molecular Biology in Jupyter Notebooks.” Biochemistry and Molecular Biology Education, vol. 50, no. 5, 2022, pp. 479–482. https://doi.org/10.1002/bmb.21676.
  • Ji, Ziwei, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. “Survey of Hallucination in Natural Language Generation.” ACM Computing Surveys, vol. 55, no. 12, article 248, Dec. 2023, pp. 1–38. https://doi.org/10.1145/3571730.
  • Qu, Y., K. Huang, M. Yin, K. Zhan, D. Liu, D. Yin, H. C. Cousins, W. A. Johnson, X. Wang, M. Shah, R. B. Altman, D. Zhou, M. Wang, and L. Cong. “CRISPR-GPT for Agentic Automation of Gene-Editing Experiments.” Nature Biomedical Engineering, 2025.
  • “Writing Is Thinking.” Nature Reviews Bioengineering, vol. 3, 2025, p. 431. https://doi.org/10.1038/s44222-025-00323-4.