What is artificial intelligence and how does it work? (Intermediate level)
Objective: The worksheet provides a basic understanding of how artificial intelligence (AI) works at an intermediate level. Learners should be able to distinguish AI applications from classic programs, understand the technical learning process, and critically question algorithms in everyday life.
Content and methods: The content covers the basic definition of artificial intelligence, the distinction between static programs and adaptive systems, the three-step model of how AI works (data collection, pattern recognition, application), and the opportunities and risks of recommendation algorithms, such as filter bubbles and data bias. This is supplemented by practical case studies and the transfer of knowledge to school application scenarios. The methods include explanatory text work, tabular classifications, visual process diagrams to illustrate technical processes, analytical case studies, and creative transfer tasks in which learners take on developer roles to design their own AI ideas for everyday life.
Competencies:
- Technical understanding: Understanding the logic of algorithms and machine learning
- Judgment: Reflection on the limitations and error susceptibility of AI systems
- Design skills: Development of innovative solutions for one's own environment (school)
- Analytical skills: Distinguishing between different phases of the learning process using practical examples
Target group and level: From grade 8 onwards
51 other teachers use this template
Target group and level
Grade 8 and above
Subjects
What is artificial intelligence and how does it work? (Intermediate level)
What is artificial intelligence?
Artificial intelligence (AI) means that computer programs can perform tasks that normally require human intelligence, such as understanding language, recognizing images, or making decisions.
Classification of applications
Analyze the digital applications listed in the table. Explain whether they are adaptive artificial intelligence or classic computer programs.
| Application | AI? Yes | AI? No | Reason |
|---|---|---|---|
| Social Media Content Moderation | |||
| Email Spam Filtering | |||
| Photo Editing Software | |||
| Online Shopping Recommendations | |||
| Voice Assistants |
Visualization of how it works
Represent the three-step process of artificial intelligence (data collection, pattern recognition, application) described in the text graphically in the form of a flowchart.
How does artificial intelligence work?
Artificial intelligence works in three simple steps. First, it collects data. The AI receives a lot of information, such as images, text, or numbers. This is called training data. Second, it learns or recognizes patterns. The AI looks for similarities and differences. For example, it recognizes that cats often have whiskers and pointed ears. Lastly, it applies or decides. When new data arrives, the AI uses its learned knowledge and makes a decision. For example, it might conclude, "This image probably shows a cat." Through these steps, AI systems develop the ability to perform tasks that typically require human intelligence, like visual perception, speech recognition, decision-making, and language translation. The effectiveness of AI depends on the quality and quantity of its training data, as well as the complexity of its algorithms, allowing it to continuously improve and adapt over time.
Lade Zeichenfeld...
Application of the three-step model
Match the practical examples below to the corresponding phases of the AI learning process.
| Description | Step |
|---|---|
| The AI receives thousands of images of dogs and cats. | |
| A new image is uploaded and the AI says, "This is a cat." | |
| The AI notices that cats often have whiskers and pointed ears. | |
| After learning, the AI provides an answer or recommendation. | |
| A computer stores lots of text, photos, and numbers. | |
| AI compares characteristics and searches for differences between animals. |
Analysis of the case study
Evaluate the case study.
- What tasks does AI perform and what specific characteristics does it need to recognize?
- Explain what user data the system needs in order to continuously improve its responses.
- Discuss the limitations of AI technology mentioned in the text.
Voice Assistants in Everyday Life
In a bustling household, Emily uses a voice assistant to manage her daily tasks, like setting reminders, playing music, and controlling smart home devices. The AI behind the voice assistant comprehends spoken language, recognizes commands, and provides relevant responses or actions. This interaction occurs seamlessly as Emily goes about her routine, whether she's in the kitchen, living room, or even driving her car. The AI functions by processing voice inputs through natural language processing (NLP) algorithms, which convert speech into text and interpret the intent using machine learning models and neural networks. These systems are trained on vast datasets to understand diverse accents, languages, and contexts, improving their accuracy over time. However, AI is not infallible; it may misunderstand commands, especially if the speech is ambiguous or the environment is noisy. Moreover, biases may emerge if the AI has been trained predominantly on data from specific demographics, leading to inaccuracies in voice recognition or interpretation for users outside those groups. Despite advancements, voice assistants sometimes struggle with complex queries or nuanced language, highlighting the ongoing challenges in AI development and the importance of continued refinement and testing to enhance user experience.
Development of optimization proposals
Create a brief optimization concept. Suggest at least two specific measures that AI could use to reduce misunderstandings in everyday life and increase the recognition rate.
Development of optimization proposals
Create your own school AI
Imagine you are a team of developers tasked with inventing artificial intelligence for your school. This AI should help students or teachers in their everyday lives. Form groups of up to 4 people.
1. Develop the idea
Think together about which problem in everyday school life your AI should solve.
Think about situations such as homework, timetables, learning, organization, or communication.
2. Plan how it will work
Now think about how your AI would work.
Every AI needs data and must be able to recognize or decide something.
3. Consider problems and improvements
AI is not perfect. Good developers always consider what could go wrong and how to make it better.
Lade Zeichenfeld...
Lösung für die Lehrkraft
| Application | AI? Yes | AI? No | Reason |
|---|---|---|---|
| Calculator | ☐ | ☑ | A calculator follows specific rules to perform math operations and does not learn from data or improve over time. |
| Music recommendations on Spotify | ☑ | ☐ | Spotify uses AI to analyze listening habits and suggest songs, learning and adapting to user preferences. |
| Autocorrect while typing | ☑ | ☐ | Autocorrect uses AI to learn from language patterns and improve suggestions, adapting to user typing habits. |
| Weather app | ☐ | ☑ | A weather app uses predefined models to predict weather based on data, without learning or adapting over time. |
| Google Maps navigation | ☑ | ☐ | Google Maps uses AI to learn from traffic patterns and provide optimal routes, adapting to real-time conditions. |
| Description | Step |
|---|---|
| The AI receives thousands of images of dogs and cats. | Collecting Data |
| A new image is uploaded and the AI says: “This is a cat.” | Applying or Deciding |
| The AI notices that cats often have whiskers and pointed ears. | Learning or Recognizing Patterns |
| After learning, the AI gives an answer or recommendation. | Applying or Deciding |
| A computer stores many texts, photos, or numbers. | Collecting Data |
| The AI compares features and looks for differences between animals. | Learning or Recognizing Patterns |
AI could be improved by making sure it understands different accents and languages better. This means using more diverse data when training the AI so it doesn't get confused when someone speaks with a different accent. Another way to improve AI is by teaching it to handle complicated questions or sentences. This can be done by having more advanced algorithms that help the AI understand context and meaning better.