The integration of artificial intelligence (AI) in education presents exciting opportunities to enrich learning experiences and prepare students for a future where these tools will play an increasingly vital role. However, two key challenges arise. First, the educational community has not yet developed a pedagogy that helps students understand and fully harness the vast potential of AI. Second, there is ongoing uncertainty about how to incorporate AI in a way that fosters, rather than undermines, the development of critical thinking and student autonomy.
We believe a compelling way to tackle these challenges is by combining hands-on science experiments with AI exploration. This approach allows students to engage with AI while simultaneously learning new scientific concepts. Science offers an ideal context to harness the power of modern generative models. Already used by researchers to accelerate their work, AI can enhance every stage of scientific inquiry—from literature review and data analysis to modeling and presenting results. In a practical experimental setting, AI not only sparks excitement but also broadens educational opportunities.
In the sections that follow, we invite you to join us in this experiment and discover what could be the interactions between students and AI, offering fresh insights into how AI's generative models can be integrated into educational practices.
An inquiry-based method boosted by AI
The inquiry-based method is based on the idea that students learn most effectively by being active and exploring on their own, rather than passively receiving information. In this method, the student uses all the tools at his or her disposal to "investigate" a phenomenon in the manner of a researcher.
Our proposal is to integrate AI into inquiry-based activities to enhance students' investigative capabilities. The aim is for the thoughtful use of AI to encourage students to delve deeper into questioning and the scientific process. With AI tools at their disposal, their potential for exploration and discovery is significantly expanded. The scientific method is strengthened as students can focus more on critical thinking and applying their knowledge.
An AI-enhanced scientific investigation could include the following steps:
Documentary Research: Students use virtual assistants to efficiently conduct comprehensive research on different aspects of the problem, gaining a deeper understanding of the topic.
Experimentation: AI guides students in selecting the best instruments and avoiding common mistakes during the experimental phase.
Data Analysis: Through dialogue with AI, students analyze their data, identify patterns or anomalies, and visualize their findings graphically.
Modeling: AI assists students in refining a theoretical model based on their observations and experimental data.
Model Study and Discussion: Students use AI to explore the model's behavior under various conditions, including boundary values, and assess its sensitivity to different parameters.
Applications and Innovation: AI’s extensive databases help students discover real-world applications and potential innovations related to their findings.
Conclusions and Communication: Students collaborate with AI to produce a clear, logical, and coherent report of their investigation.
Sharing: Together, the student and AI determine the most effective way to structure and share the findings.
The educational objective of this session is to show students how AI can enhance their research, help them organize their work, initiate productive dialogues, and recognize both the strengths and limitations of these tools. Ultimately, the goal is to help students learn how to leverage AI to its fullest potential.
Experimental Focus and Practical Implementation
In this article, we propose a scientific investigation on a physical phenomenon that is straightforward for students to implement, yet complex enough for AI to play a meaningful role. This practical work is designed for 11th grade students specializing in "Physics-Chemistry" or "Engineering Sciences." The complexity of the chosen phenomenon deliberately exceeds the program's typical expectations, aiming to encourage students to seek AI assistance and push beyond their usual limits.
The challenge we present to students is: “By combining an experimental approach and AI, use your smartphone’s accelerometer to study the mechanics of a vertical elastic pendulum and produce a scientific publication.”
The publication should include the essential components of scientific research: literature review, experimental setup, data analysis, modeling and theoretical connections, model study and discussion, applications, and communication.
To carry out the experiment, students are provided with a 40 cm elastic band to serve as a spring, along with a plastic bag to hold the smartphone. This activity can be conducted in the laboratory or classroom, provided suitable spaces are available to hang the pendulums.
For AI support, we use ChatGPT in this article, but any AI system can deliver similar results, and we encourage students to select the platform they prefer. As AI tools are evolving rapidly, additional features may become available. In practice, students can access AI via classroom computers, tablets, or smartphones with the ChatGPT app.
Calibrating the AI tool
Before starting an experimental study, you need to calibrate your instruments. The same goes for AI. This process is called prompting . Prompts are the instructions or questions provided by the user to guide the AI in its response. In other words, a prompt is the starting point that tells the AI what the user wants to get in terms of information or help.
Prompts are used to:
Guiding responses : A prompt helps the AI understand the context and purpose of the question. For example, a specific and well-formulated prompt helps the AI generate a more relevant and targeted response.
Set tone and structure : The prompt type can tell the AI what style of response to expect (formal, informative, creative, concise, etc.). A prompt asking for a simple explanation for a high school student, for example, guides the AI toward a more educational and accessible tone.
Specify actions or formats : Some prompts can ask the AI to structure the response in a specific way (e.g., as a list, paragraph, or worked example). This allows you to receive a response that matches a particular format or requirement.
Customize results : Prompts allow you to tailor the response to specific needs, such as asking for an explanation tailored to a certain level of knowledge or a specific domain (science, history, technology, etc.).
To begin work with a generative model, it is therefore important to write an initial prompt that defines the context, the objective of the research, the type of answers that we want to obtain, and the level of complexity of the answers. It is essential that the student and his group, with the help of the teacher, define one or more introductory prompts adapted to the research and the level of the students. The risk if this introductory phase is not followed is that the AI gives results that are too complex or poorly adapted to the level of the students.
By clicking on the > tab below, you can view the prompt we used for the rest of the exercise.
See the Dialogue with ChatGPT
Documentary research
Document research using artificial intelligence is the simplest and most common use of generative models. However, within the framework of the scientific investigation method, AI brings important educational advantages.
The aim of the literature review is to provide students with the theoretical foundations needed to understand the swinging pendulum, exploring key concepts such as oscillations, damping, and mathematical modeling. It also prepares them to use modern measurement tools, such as smartphone sensors, and to lay the foundations for the analysis and modeling of experimental data. Finally, it encourages structured scientific thinking before moving on to experimental manipulation.
AI is interesting pedagogically in this context because it allows rapid access to precise information adapted to their specific questions through an interactive dialogue. It guides them in their thinking by suggesting relevant avenues and clarifying complex concepts. AI personalizes learning by adapting to the students' level of understanding, while synthesizing information to allow them to focus on the essentials. By facilitating understanding and guiding research, AI allows students to structure their scientific approach in a more fluid and efficient manner.
By opening the tab > below, you will discover the beginning of a conversation between a student and an AI, which allowed the student to circumscribe the subject and obtain initial hypotheses on the mathematical model of the acceleration of the pendulum.
See the Dialogue with Chat GPT
Experimental work
Once the documentary research is completed, the student and his group move on to the experimental manipulations. To succeed in this phase, several elements must be brought together: how to set up the experimental device, the safety rules to follow, record the data and export them for analysis. These aspects can be submitted to the AI to obtain suggestions, adjusted according to the context of the experiment. Below we propose several questions that the student can ask the AI.
It is important to note that AI is not always the ideal solution, which is a valuable learning experience for students. For example, the explanations provided by AI on the experimental setup may lack precision. In this case, the student will have to experiment to find a simple and effective solution on their own. In addition, AI, like ChatGPT, does not always know certain technical specificities, such as the existence of a direct measurement of the norm on FizziQ, which students will discover along the way.
To view the AI chat, tap the > tab below.
See the Dialogue with ChatGPT
After verifying that the generated graph has enough oscillations, the data is exported to a CSV file. The experimental phase is therefore complete, we can now return to using AI to analyze the data!
Data analysis
Data analysis aims to examine the recorded experimental information to extract key elements and draw conclusions about the physical characteristics of the observed motion. This includes removing unnecessary segments to clean the data, smoothing the data to reduce noise, and identifying features such as the period of the motion or the damping of oscillations.
Modern generative models allow not only in-depth graphical or statistical analysis, but also direct data reprocessing through a natural language interface, which is more intuitive for the student. The student can of course ask for help to perform these operations or understand their usefulness.
Quickly, the student will be able to:
Visualize data : To obtain a graphical representation that makes it easier to understand trends.
Clean the data : By eliminating anomalies or outliers that could distort the analysis.
Smooth data : To smooth out random fluctuations and better see major trends.
Calculate and interpret characteristics : How to determine the period of oscillations and evaluate the effectiveness of damping.
Export data : To use in other software or share.
In our dialog below, we analyzed and cleaned the data coming from the table exported in the experimental part. This analysis revealed important aspects of the behavior of acceleration over time and allowed us to detect invariants in the motion.
Open the > tab to discover the dialogue between the student and the AI
See the Dialogue with ChatGPT
Modeling
In scientific reasoning, the relationship between theory, model, and data can be articulated in several ways, each reflecting a different approach to research. A direct method is to start with theory and verify that experimental data confirm the mathematical equations derived from it. It is also possible to construct a mathematical model that seems to explain the data well, in order to guide theoretical thinking in a specific direction. Both of these approaches can be facilitated by the use of artificial intelligence.
In our example, during their documentary research, the students were directed towards a periodic sinusoidal evolution of the acceleration. By analyzing the curves, they also observed that, although this evolution is periodic, the amplitude of the oscillations seems to decrease with time. These observations are transmitted to the AI to try to determine a model that explains the data, and to obtain indications on the theoretical reasons explaining this model.
To view the AI chat, open the > tab below.
See the Discussion with ChatGPT
We see through these exchanges that AI can adapt to the level of the person it is exchanging with. For a first-year student, the explanation is much clearer. It can be even simpler if necessary.
Model analysis
The purpose of the model study is to analyze the behavior of the determined theoretical model. In the context of our study, the student seeks to understand how different parameters of the system – such as the damping or elasticity of an elastic band – influence the oscillations of an elastic pendulum . Some of these parameters can only be discovered after several round trips in the questioning, a process that contributes to the student's learning.
This study often requires the creation of a computer program to generate curves and facilitate the exploration of parameter variations. AI is an excellent programming assistant and can help the student in this creation process. The tool can also perform these analyses online, using natural language.
To discover the exchange between the student and the AI, open the dialog box > below:
See the Dialogue with ChatGPT
Communication
Writing a research paper is probably the best-known part of generative model applications. AI can help a student write a well-structured, coherent, and audience-friendly research paper by providing support throughout the writing process.
1. Structuring the document
2. Clarity and fluidity
3. Advice on adapting to the public
4. Reminder of things not to forget
5. Consistency in argumentation and scientific reasoning
6. Final check and proofreading
List of questions for ChatGPT
Sharing
Students will likely be interested in another aspect of research which is its dissemination. AI can greatly facilitate the dissemination of a research paper and optimize its online visibility through SEO (Search Engine Optimization) techniques.
AI will be able to guide the student for these different tasks:
SEO Optimization : AI suggests relevant keywords, writes optimized title tags and meta descriptions, and checks keyword density to improve SEO on Google.
Creating titles and subtitles : She reformulates titles and subtitles to make them more catchy and optimized for SEO.
Summary writing : AI produces impactful and suitable summaries for social networks or newsletters, thus attracting readers.
Content for social networks : It generates specific posts for each platform (Twitter, LinkedIn, Facebook), ensuring optimal distribution.
This part of the work also allows the student to reflect on how search engines provide their answers.
List of questions to ChatGPT
Conclusion
In this article, we have proposed a practical and motivating approach to introduce students to the vast possibilities offered by generative models, by integrating artificial intelligence (AI) into scientific practical work (TP).
This method seems particularly suitable for introducing students to the use of AI in a school setting. By integrating it into an experimental research context, it encourages an open and exploratory approach, where students can discover and use this tool independently. The sciences offer a particularly fertile ground for this approach, because AI can be used at different stages: documentary research, data manipulation and visualization, theoretical analysis, program creation, as well as communication and dissemination of results. The natural language interface simplifies scientific work, making complex concepts more accessible and allowing students to manipulate them intuitively. In addition, the ability of AI models to adapt to students' objectives and level, thanks to personalized instructions (prompts), makes it an ideal tool for a wide variety of users.
We welcome your feedback on this experience. AI should not be seen as a definitive solution, but rather as a catalyst for reflection and innovation. It helps students not only master science, but also develop essential digital skills for the 21st century. The challenge now is to find a balance between exploiting this promising technology and maintaining the intellectual autonomy of learners, in order to develop critical minds capable of using these tools in an informed and responsible manner.
Comentarios