My First Time Trying to Explain How AI Learns – TELC B1.2 Writing Challenge

Okay, so here I am, almost six months in Berlin. The language is… a lot. And my German isn’t exactly fluent yet. I’m taking this TELC B1.2 writing course as part of the integration program, and honestly, it feels like climbing Mount Everest sometimes. My task today was to write an explanation about how AI learns from data – specifically for the exam. It sounded simple enough on paper, but actually putting it into words in German felt incredibly tricky.

The Brief: Explain it to Frau Schmidt

The prompt said something like: “Erklären Sie, wie künstliche Intelligenz aus Daten lernt.” (Explain how artificial intelligence learns from data.) I immediately panicked. I knew something about this because I’ve been reading a lot about Google Assistant and Siri – I ask them all the time for directions (“Wie komme ich zum Hauptbahnhof?” – How do I get to the main train station?) But actually explaining it clearly, in a way someone like my colleague, Frau Schmidt (she’s always so patient), would understand… that was another thing entirely.

My Initial Attempt – A Complete Mess!

I started writing and, well, let’s just say it wasn’t pretty. I used some really complicated vocabulary that sounded amazing in my head, but then I realised I’d completely lost the thread. I wrote about “Algorithmen” (algorithms) and “neuronale Netze” (neural networks) without actually explaining how they learn. I felt so frustrated!

Here’s a little bit of what I produced – don’t judge too harshly:

“Künstliche Intelligenz nutzt komplexe Algorithmen und neuronale Netze, um aus Daten Muster zu erkennen. Die Datengrundlage ist essentiell für die Lernfähigkeit der KI.” (Artificial intelligence uses complex algorithms and neural networks to recognize patterns from data. The data foundation is essential for the learning ability of AI.)

It sounded… pretentious. And completely useless. Frau Schmidt just blinked at me, politely saying, “Das ist… sehr technisch.” (That’s… very technical).

A Real Conversation – Learning From Mistakes

I needed a different approach. I talked to my flatmate, Steven, who’s actually pretty good with computers. He explained it to me like this: “Stell dir vor,” he said, “die KI bekommt Millionen von Bildern von Katzen. Sie schaut sich jeden einzelnen an und merkt, was eine Katze ausmacht – die Ohren, die Schnurrhaare, die Form des Körpers. Nach einer Weile kann sie selbstständig eine Katze erkennen, auch wenn es eine neue Katze ist, die sie noch nie gesehen hat.” (Imagine,’ he said, ‘the AI gets millions of pictures of cats. It looks at each one individually and notices what a cat is – the ears, the whiskers, the shape of the body. After a while, it can independently recognize a cat, even if it’s a new cat that it has never seen before.)

“Und das ist ‘Lernen aus Daten’!” (And that’s ‘learning from data!’) He gave me an example: “Wenn ich Google Maps frage, wo der nächste Bäcker ist, dann lernt Google, welche Geschäfte Bäcker sind und wie viele es gibt – basierend auf den Suchanfragen anderer Leute.” (When I ask Google Maps where the nearest bakery is, then Google learns which shops are bakeries and how many there are – based on the searches of other people.)

Suddenly, it clicked! It wasn’t about the complicated technical terms. It was about examples.

Breaking it Down for the Exam – TELC Style

Okay, so thinking about what I need to do for the exam… I realized the TELC B1.2 writing focuses on explaining things clearly and using appropriate vocabulary. My explanation needs to be understandable for someone who isn’t an expert in computer science.

Here’s how I’m planning to approach it, based on Steven’s explanation:

1. Start with a simple definition: “Künstliche Intelligenz lernt, indem sie viele Beispiele bekommt.” (Artificial intelligence learns by getting many examples.)

2. Give a clear example: “Denken Sie an Google Assistant. Wenn Sie ihm sagen ‘Senden Sie eine SMS an meine Mutter’, dann lernt er, wie man eine SMS schreibt.” (Think about Google Assistant. When you tell him ‘Send an SMS to my mother’, then he learns how to write an SMS.)

3. Keep it concise: Don’t get bogged down in details. Focus on the core concept.

Practical German Phrases for Your Own Writing

Here are some useful phrases I’ve picked up that might be helpful:

• “Basierend auf…” (Based on…) – Useful when explaining the source of information.
• “Das bedeutet, dass…” (That means that…) – For connecting ideas and providing explanations.
• “Zum Beispiel…” (For example…) – Perfect for illustrating your point with real-world examples.

Feeling a Bit Better Now…

Writing this has actually helped me understand the concept better. It’s not about knowing all the fancy words, it’s about communicating clearly in German. And I think, with practice and more conversations like these, I’m going to get there – eventually! Jetzt muss ich mich auf die Hausaufgaben konzentrieren (Now I need to focus on my homework!).

—