At SARSEF we recognize the growing interest in utilizing Artificial Intelligence (AI) technology, including Large Language Models (LLMs), Machine Learning (ML) and Generative AI for research projects. As such, we provide these guidelines for you, the researching student, as well as any supporting adults to responsibly engage with AI technologies while maintaining project integrity and upholding ethical standards. Familiarize yourself with our new AI policy for Science and Engineering Fair by reading the guidelines below or visiting our Scientific Integrity page.
Responsible Use of Generative AI Tools
Student Ownership: AI tools may be used to support your learning, not replace it. All projects submitted to SARSEF must be your original work. The Society for Science added this statement about AI use to the Eligibility/Limitations section of the ISEF rulebook: “Artificial Intelligence may be used as a project resource but must be cited and given proper acknowledgements. All materials presented must be in the researcher’s own words.” You may not use AI to write your research plan, abstract, paper, or poster. AI may only be used as a resource and must be cited. While generative AI tools are allowed, its use is optional, not mandatory.
Purposeful Engagement: You can use AI tools as aids for exploring topics, searching for information, clarifying complex texts, or checking grammar. However, you are strongly encouraged to read primary and secondary sources, analyze evidence, and make independent choices in your writing and presentations. Personal details about you or any research subjects should not be shared with AI. Adult
Sponsor Involvement: Adult sponsors are highly encouraged to collaborate with students to demonstrate genuine engagement with the project and keep students apprised of how and when to use AI.
Understanding AI and LLMs
LLMs like Chat GPT generate human-like text based on input. They operate with a knowledge cutoff (i.e. any new information, events, or developments after the tool was trained will not be accessible) and therefore may lack up-to-date information. They can provide general information but might not always be accurate or context aware.
When using AI, avoid sharing personal details or any personal identifying data about you or any research subjects. Information shared on AI runs the risk of being shared publicly.
Treat AI as a tool, not a confidant.
Citation and Verification
- Verify AI-generated information from reliable sources. AI-generated information may be used as a starting point for research.
- Ask AI for its sources and properly cite generative AI as references.
- You must document all interactions with AI, including the questions you ask and the responses received. In addition to citation in references, we strongly recommend a copy of the transcript when utilizing any GenAI interface. This may be uploaded as a document in the “lab journal excerpts” field on the registration. Make sure to include the name of the AI system used and the date utilized.
- You should maintain a logbook to ensure transparency in the research process.
- Remember, AI is a powerful tool, but it’s essential to use it responsibly and ethically. Failure to cite GenAI use may be grounds for disqualification.
Addressing Bias
- AI may inherit biases from its training data. Critically assess AI-generated content for potential biases.
- You should address identified biases in your project and work to mitigate any bias wherever possible.
Ethical Considerations
- You should think about the ethical implications of using AI in your research. Uphold ethical principles in your work and discussions involving AI.
- Employ AI only for legitimate purposes (see chart on last page for examples). Avoid generating false information, deception, or harm.
- Use AI tools as a supplement but rely on your independent analysis and research.
- Do not claim AI-generated essays as your own original work. Ethical concerns arise when the AI program produces essays that are claimed to be the original work of a person.
- Remember that AI is a tool to enhance your research, not a replacement for critical thinking and creativity.
AI Limitations
- LLMs such as ChatGPT may not always understand complex or context-specific questions.
- To receive accurate responses, start with specific, clear inquiries.
- Refine your questions to target the information you need and be patient and persistent in refining until you obtain accurate answers.
- Generative AI is subject to “hallucinations,” where outputs sound confident and plausible, but are fabricated and factually incorrect.
- Remember “garbage in = garbage out.”
Additional Resources
AI vs. Machine Learning vs. Deep Learning vs. Neural Networks: What’s the difference?
Video: How Large Language Models Work
Harvard Business School: Student Use Cases for AI
ChatGPT: An Introduction for Educators
Statement from the International Baccalaureate about Artificial Intelligence in Education
Ethical AI for Teaching and Learning
AI Tools
GenAI Use and Student Research Projects
ALL GenAI use must be documented and cited; failing to do so may result in disqualification.
| Research Assistance AI tools may help students gather background information about their topic quickly. This may aid them in developing a strong knowledge base. All information and sources should be verified. | Idea Generation AI tools may assist students in brainstorming project ideas and suggesting research questions, hypotheses, or experimental design. | Data Analysis The use of AI may help guide students on data analysis techniques, statistical tests, or software tools for processing and interpreting results. Students should draw their own conclusions from the evidence presented. |
| Experimental Procedure AI can assist in offering suggestions to improve accuracy, control variables, and ensure safety, but it cannot write the research plan. | Explanation and Clarification Students can ask AI to explain complex scientific concepts, theories, or equations related to their projects, aiding their understanding and communication of the topic. | Source Recommendation AI can provide suggestions for credible sources, articles, books, or research papers. The student is expected to review the sources and create their own bibliography. |
| Figure Creation Students can ask for guidance on creating effective visual representations of their data, such as graphs, charts, or diagrams to enhance the presentation of their results. All figures must be properly cited including the method used to create student images/figures. | Ethical Considerations AI can help students evaluate the ethical aspects of their projects, including considerations of human or animal subjects, environmental impact, or data privacy. Students should not share personal information or details about themselves or any research subject with AI. | Presentation and Communication Students can seek tips on creating engaging project presentations, posters, or oral explanations for the science fair judges and attendees. However, students may not use AI to write the presentation for them. |
This information has been adapted from the following sources: SSTFI, GARSEF, DMWSEF, College Board, and BRSF.