Das Schreibseminar.

Alex D via Midjourney

At present, you can usually unmask the creator as artificial intelligence in most images and videos by counting the people’s fingers. It’s not certain whether AI truly struggles to depict fingers and toes accurately or if it’s a trick designed to give us a false sense of security. Either way, if images look too good to be true, they’re probably creations of non-organic neural networks.

Texts written by AI have been around for quite some time. They’re continually improving and, especially in marketing, are now hardly distinguishable from those composed by soulless product manager aspirants. The human rowers in the corporate galleys produce pitiful work compared with well-prompted LLMs.

Does a promotional text feel stale and formulaic to you? Then it’s probably by a human. By using so-called “humanizer” software, AI-generated texts can be spiced up with slight irregularities and linguistic quirks. That already tastes better than the output of fatalistic wetware copywriters. Such “humanized” texts are also harder for software intended to detect AI authors to spot.

How, then, can we identify AI-generated texts? What’s the equivalent of counting fingers? I propose the following method:

1. Is the text boring, formal, polished to a gaudy gleam? Then it could be from either a human or a machine—it makes no difference.  

2. Is the text professional but contains subtle oddities? Then it might be from a worn-out copywriter or an excellent machine. That’s the future.  

3. Do you find surprising twists, original sentence structures, and striking neologisms? Then it’s definitely by a human—or possibly by an AI that was trained by a human using extensive examples of their own style.  

4. Is the text rife with hallucinations? In 2025, it’s almost certainly by a substance-loving human or the free version of what is otherwise a powerful AI.  

5. Could a text like the one at hand be attributed to an AI? Not yet.

So what? Some human writers scratch away with fountain pens on paper. Others clatter on computer keyboards. Others dictate, and it’s transcribed later. And many already engage in a fruitful dialogue with a large language model to create interesting writing—just as numerous significant novels have come from the collaboration of a purposeful author and a capable editor.

Conclusion: Forget about counting fingers. If a text fulfills its purpose, it’s good—regardless of whether it was written by a human, a machine, or both.