Dein erstes KI-Experiment
Your First AI Experiment 🧪
You've learned what AI is, how machines learn, how your phone uses AI, and how chatbots work. Now it's time to roll up your sleeves and try it yourself.
This lesson guides you through four free experiments. You don't need any technical skills - just curiosity and a phone or computer.
🔬 Experiment 1: Image Recognition with Google Lens
What you'll try: Show AI a photo and watch it identify what's in it.
How to do it:
- Open the Google app on your phone (or visit lens.google.com)
- Tap the camera icon (Google Lens)
- Point your camera at any object nearby - a plant, a book, a piece of food, a landmark
- Watch as the AI identifies the object and gives you information about it
Try these challenges:
- Point it at a houseplant - can it tell you the species?
- Point it at text in another language - can it translate it live?
- Point it at a product - can it find where to buy it?
Google Lens processes over 12 billion visual searches every month. It can identify more than 15,000 plant species and recognise thousands of dog and cat breeds from a single photo.
When you tried Google Lens, did it get everything right? Where did it struggle? AI is excellent at identifying common objects but can stumble on unusual angles, poor lighting, or rare items. What does this tell you about how AI "sees" the world compared to how humans see it?
✍️ Experiment 2: Text Generation with ChatGPT
What you'll try: Have a conversation with an AI chatbot and see how it responds.
How to do it:
- Visit chat.openai.com and create a free account (or use any free chatbot)
- Start with a simple question: "Explain why the sky is blue as if I'm 8 years old"
- Then try something creative: "Write a short story about a cat who becomes a detective"
Try these prompts:
- "Summarise the plot of my favourite film in exactly 3 sentences" (name the film)
- "Give me 5 unusual facts about the country I was born in" (name the country)
- "Write a polite email asking my neighbour to turn down their music"
Now test its limits:
- Ask about a very recent news event - does it know about it?
- Ask it to solve a tricky maths word problem - does it get it right?
- Ask it the same question twice - do you get identical answers?
Why might ChatGPT give a different answer when you ask the same question twice?
🎨 Experiment 3: AI Image Generation
What you'll try: Describe a picture in words and watch AI create it from scratch.
How to do it:
- Visit bing.com/images/create (free with a Microsoft account)
- Type a description of an image you'd like to see
- Wait a few seconds and see what the AI generates
Try these descriptions:
- "A cosy bookshop on a rainy London street at night, watercolour style"
- "A golden retriever wearing a tiny astronaut helmet floating in space"
- "A futuristic city built entirely of glass, with flying bicycles"
Things to notice:
- How well does it follow your description?
- Look closely at hands and text in the image - what do you spot?
- Try adding "in the style of a pencil sketch" - how does the output change?
AI image generators occasionally produce odd results - extra fingers, warped text, or strange perspectives. These quirks happen because the AI learned from millions of images but doesn't truly "understand" anatomy or physics. It's pattern matching, not comprehension.
🌐 Experiment 4: AI Translation with Your Voice
What you'll try: Speak in one language and hear AI translate it into another - in real time.
How to do it:
- Open Google Translate on your phone or visit translate.google.com
- Tap the microphone icon
- Speak a sentence in English
- Select a target language (try Spanish, French, Japanese, or Arabic)
- Tap the speaker icon on the translated text to hear it spoken aloud
Try these challenges:
- Translate a sentence, then translate it back to English. Is the meaning preserved?
- Try a sentence with slang or an idiom: "It's raining cats and dogs" - what happens?
- Try speaking quickly or with background noise - how well does it cope?
What might happen when you translate an English idiom like 'break a leg' into another language using AI?
📓 Reflection: What Surprised You?
Now that you've completed your experiments, take a moment to reflect.
Think about your four experiments. Which AI tool impressed you the most? Which one made a mistake that surprised you? Did anything feel slightly unsettling or magical? Understanding your own reactions to AI is just as important as understanding the technology itself. Your instincts about what feels right and wrong will guide you as AI becomes more powerful.
Consider these questions:
- Which AI felt the most "intelligent" to you?
- Where did AI clearly fail or get confused?
- Would you trust these tools for something important - like homework, a work email, or medical advice?
- How do you feel knowing these tools are available to everyone, for free?
🔑 Key Takeaways
- Google Lens can identify objects, plants, text, and products from photos
- ChatGPT generates human-like text but can give different answers each time
- AI image generators create pictures from text descriptions but struggle with details like hands and text
- AI translation is impressive for common phrases but can stumble on idioms and slang
- AI tools are powerful but imperfect - always verify important results
- Your own critical thinking is the most important tool of all
After trying these AI experiments, what is the most important habit to develop?
🎉 Congratulations!
You've completed Level 1: AI Seeds - the absolute beginner's guide to artificial intelligence. You now understand:
- ✅ What AI is and isn't
- ✅ How machines learn from data
- ✅ The AI hidden in your smartphone
- ✅ How chatbots like ChatGPT work
- ✅ Hands-on experience with real AI tools
You're no longer a complete beginner. You're someone who understands the basics, has tried the tools, and can think critically about what AI can and cannot do.
Welcome to the world of AI. 🌱