Training AI: AI & Machine Learning
Scientific Reasoning & Ethical Awareness
Resources & Downloads
Download the PDF versions of this scenario or the associated attachments.
Target Group: 13 - 16 y.o.
Activity Duration: 4-5 lessons (45 min each)
Key Learning Goals:
- Scientific Reasoning: Deepen understanding of science concepts (e.g., classifying rocks, plants) using AI contexts.
- Critical Thinking: Analyze AI performance, identify limitations, and evaluate if AI can replace human judgment.
- Digital Literacy: Use tools like Teachable Machine to train models and investigate data quality.
- Ethical Awareness: Reflect on the implications of using AI in scientific analysis and decision-making.
Learning Outcomes
Students will be able to:
KNOWLEDGE & UNDERSTANDING:
- Describe machine learning concepts (training, testing, datasets) and AI applications in science.
- Identify sources of error and bias in AI models.
- Recognize ethical considerations like trust and responsibility in automated decision-making.
SKILLS & ABILITIES:
- Collect and label datasets for training AI models.
- Use Teachable Machine to create and evaluate predictive models.
- Collaborate in teams to design experiments and analyze results.
ATTITUDES & VALUES:
- Appreciate the synergy between human insight and technological tools.
- Demonstrate openness to diverse perspectives in scientific discussions.
- Show curiosity about the intersection of AI and natural sciences.
European Dimension / Erasmus+ Connection
- Digital Responsibility: Encouraging ethical use of data and AI tools.
- Cross-Border Collaboration: Sharing datasets and results with European partner schools.
- Scientific Citizenship: Engaging students in real-world scientific debates and inquiries.
1. Resources and Tools
Digital Tools:
- AI Training: Teachable Machine (Google).
- Collaboration: eTwinning/TwinSpace, Padlet.
- Presentation: Canva, PowerPoint.
Materials:
- Computers/Tablets with cameras.
- Physical objects for classification (e.g., different types of rocks, leaves, or printed images).
- Worksheets (Ethics & Debate, Evaluation forms).
2. Working Methods
- Scientific Inquiry: Formulating hypotheses, collecting data, and analyzing results.
- Project-Based Learning: Creating a functional AI model to solve a classification problem.
- Debate: Structured discussion on the ethical implications of AI in science.
- Collaborative Learning: Working in teams to build datasets and train models.
Activity Overview
| Phase | Duration | Activity | Description |
|---|---|---|---|
| Intro | 20 min | AI in Science | Video & Discussion: "How AI is changing science." Introduction to the concept of Machine Learning as a tool for scientists. |
| Activity 1 | 45 min | The Human Algorithm | Manual Classification: Students classify objects (e.g., rocks, plants) using their own criteria to understand the "human" process before automation. |
| Activity 2 | 45 min | Collecting Data | Fieldwork: Teams collect and label images to create a dataset. Discussion on data quality and diversity. |
| Activity 3 | 45 min | Training the Model | Teachable Machine: Training the AI with the collected dataset. Testing accuracy and identifying errors. |
| Activity 4 | 45 min | Ethics & Debate | Critical Analysis: "Can AI replace lab scientists?" Debate and reflection on the role of human judgment. |
3. Introduction and Motivation
AI in Science
Goal: Contextualize AI as a scientific tool.
- Watch: Video on AI in fields like biology or astronomy.
- Discuss: "How does a computer 'know' what a cancer cell looks like?" (It learns from examples).
4. Research and Learning
Activity 1: The Human Algorithm
Task: Students act as the "algorithm."
- Given a set of objects (e.g., images of sedimentary vs. igneous rocks), students create their own rules for classification.
- Reflection: How did you decide? Was it easy? Where did you disagree?
Activity 2: Collecting Data
Task: Building the Dataset.
- Teams capture photos of their assigned category.
- Challenge: Ensure diversity (lighting, angles) to prevent bias.
- Output: A labeled dataset ready for training.
5. Creative Application
Activity 3: Training the Model
Tool: Teachable Machine.
- Train: Upload datasets into classes.
- Test: Present new objects to the AI.
- Analyze: Did it work? If it failed, why? (e.g., background noise, insufficient data).
6. Reflection and Evaluation
Activity 4: Ethics & Debate
Topic: "Can AI replace lab scientists?"
- Debate: Teams argue for and against the reliability of AI in critical scientific tasks.
- Reflection: Students post their final thoughts on a Padlet board.
- Evaluation: Complete the self-evaluation and peer feedback forms.
Co-founded by the Erasmus+ Programme of the European Union • 2024-1-HR01-KA220-SCH-000254267
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License.