Tools like ChatGPT and Claude can feel like magic: you type a question, and a thoughtful, well-written answer appears in seconds. Here's what's happening behind the scenes.

They Predict, They Don't "Know"

These tools are built on large language models, systems trained on enormous amounts of text to learn the patterns of language: which words and ideas tend to follow which others. When you ask a question, the model isn't looking up a stored answer. It's generating a response one piece at a time, predicting what should come next based on everything it learned during training.

That's a subtle but important distinction. It means these tools are exceptional at producing fluent, well-structured, plausible-sounding text. But "plausible-sounding" and "correct" aren't always the same thing.

What They're Good At

  • Drafting and rewriting: turning a rough idea into a structured first draft
  • Summarising long documents or explaining a complex topic simply
  • Brainstorming: generating options quickly when you're stuck
  • Editing: tightening wordy sentences, checking tone and clarity

Where They Fall Short

These tools can sometimes state incorrect information with total confidence (often called "hallucinating"). They don't automatically know about very recent events unless the tool is specifically connected to the internet. And they don't know you (your actual achievements, your voice, your context) unless you give it to them clearly.

The Practical Takeaway

Use these tools as a fast, capable first-draft partner, never as the final word. Always read critically, fact-check anything that matters, and make sure the final result still sounds like you. That's the difference between using AI well and letting it use you.