Screenshot

A screenshot is a photo of the screen that the device itself takes. So you don’t hold a camera up to the monitor. Instead, the computer or phone saves exactly the pixels it is currently displaying. The result is an ordinary image file, usually in PNG or JPG format. In German this is also called a Bildschirmfoto. Because the image comes directly from the device, it is razor-sharp and shows no reflections or crooked edges.

The screenshot as evidence and as an explanatory aid

Much of what appears on screens is fleeting. A message gets deleted, a price changes, a post disappears. A screenshot freezes this state. That’s why screenshots turn up in messages, court proceedings, and disputes as evidence.

However, this kind of evidence is only reliable to a limited extent. An image can be altered with simple programs, and fake chat logs regularly circulate online. Experts therefore check details like timestamps, fonts, or line spacing. A screenshot alone is rarely considered strong evidence in court.

The second major use is explanation. Anyone reporting a bug saves long descriptions by using an image instead. Tutorials, homework help, and technical support all rely heavily on screenshots. An image immediately shows what would sound cumbersome in words.

What happens inside the device when you take one

The screen content exists inside the device as a grid of pixels, stored in a dedicated memory area. When you take a screenshot, the operating system copies this area and writes it to a file. That’s why it happens so quickly and uses barely any processing power.

It is triggered using fixed key combinations. Windows has the Print Screen key, while Apple computers use the combination of the Command key, Shift key, and a number. On a phone, you usually press the power button and volume button at the same time. Many systems also allow you to select just a portion of the screen.

A common misconception: a screenshot does not contain text in the true sense. It only contains pixels that look like text to our eyes. A program first has to recognize the letters, a process called text recognition or OCR. The screenshot should also be distinguished from screen recording, which produces a video instead of a still image. Some apps, such as streaming services or banking apps, block screenshots for legal and security reasons.

Screenshots as fuel for AI assistants

In everyday life you encounter screenshots constantly: as a forwarded chat log, as proof of an order, as a clip from a video. On social networks, entire discussions are shared as images because the original is behind a paywall or has been deleted.

What’s new is the role screenshots play in artificial intelligence. Modern chatbots process not only text but also images. You send a screenshot of an error message or a math problem and get an explanation back. Such systems are called multimodal because they can understand multiple types of input.

So-called computer agents go a step further still. These programs continuously look at screenshots of the screen to figure out where buttons are located. They then control the mouse and keyboard themselves and carry out tasks such as filling out forms. This approach has regularly appeared in reports about Microsoft, Google, or OpenAI since 2024. Privacy is a point of critical debate here, since a screenshot often reveals more than was intended.