How can I tell if something I'm reading was written by AI instead of a human?

You can't reliably tell if something was written by AI, and anyone claiming otherwise — including AI detection tools — is overselling their accuracy, but there are patterns that often signal AI-generated text. The most common tells are repetitive sentence structures, a complete absence of personal anecdotes, and a tendency to hedge every claim with phrases like 'it's important to note' or 'in today's digital landscape.' AI also struggles with specific, verifiable details. It'll say 'many companies use AI' but won't name which companies or cite real numbers. Here's a concrete example: I once read a blog post about coffee brewing that claimed 'the optimal water temperature is between 195 and 205 degrees Fahrenheit.' That's technically correct, but the post never mentioned the author's own brewing mistakes, never referenced a specific coffee bean, and never shared a single personal experience. It read like a Wikipedia summary. A human coffee nerd would've told you about the time they scorched their beans with water that was too hot. That said, detection tools like GPTZero or Originality.ai are unreliable. A 2023 study from Stanford found that AI detectors misclassified human-written text as AI-generated up to 20% of the time, especially for non-native English speakers. So don't use them to accuse anyone. Instead, focus on whether the content is actually useful. If it's vague, generic, and lacks original insight, it doesn't matter who wrote it. For a deeper dive, see our guide on [AI writing sounds too formal](/blog/ai-writing-sounds-too-formal). **Related**: Are AI detection tools accurate? | What are the signs of AI-written text?