Artificial intelligence "AI, a picture for the birthday card, please!"
Artificial intelligence provides support in everyday life and creates visual content quickly and easily. With the following image generation models, you can turn your ideas into images in no time at all.
No time? blue News summarizes for you
- Image AIs build an image pixel by pixel and are specialized for different applications.
- In order to generate the desired images, it is worth structuring the prompt.
- Swisscom offers an online guide and free webinars on AI.
Not everyone knows that artificial intelligence (AI) as an image program supports the creation of visual content. Yet AI image programs are an important little helper in everyday life. They generate images and graphics or even edit them - including funny effects, of course.
AI: a little helper with many qualities
AIs generate images or logos and even edit drawings. If you need a new logo for your club or an invitation card, simply commission an image generation program.
Various image generation programs are capable of converting sketches into a desired style or redesigning areas of a photo: The gray background in the family portrait is transformed into a sunset with a single instruction - a prompt.
How image AIs work
Image generation models generate images from text by building up an image pixel by pixel. In doing so, they are guided by the prompt. A prompt refers to the input you make in the text field.
If you want to try out AI, you can choose between various programs such as Midjourney, Dall-E, Adobe Firefly or Stable Diffusion. To ensure that the graphics, for example for the birthday card, meet your expectations, it is worth using a prompt that is as accurate as possible.
Image generation: the right prompt
A good prompt provides specifications for specific content, visual style, such as colors and technique, or the desired mood in the image.
«To achieve a good result, avoid unnecessary filler words and give instructions that are as specific as possible. References from photography and art can help here.»
Media expert and trainer at Swisscom.
For the birthday card, the prompt could be: "Generate an impressionist-style picture of a ladybug painting a picture at a table, warm colors." This is what the result looks like:
Some AIs work in German, others require input in English. Searching for technical terms from the visual arts is also helpful. The AI is trained in these terms and uses them to generate an image in the desired style. For example: pop art aesthetics, deep perspective, color palette with warm tones.
Swisscom Campus as a digital companion
If you are curious, try out an AI image program online for free and discover various functions and effects for creating images. You don't need any previous knowledge of graphics.
Tips and examples of AI applications can be found on Swisscom Campus, as well as information on other digital topics. If you would like to learn more about artificial intelligence, register now for Swisscom's online course, which is free of charge.
These are the best-known AI image generators
- Midjourney is characterized by high-quality results and little effort in prompting.
- Dall-E from OpenAI is integrated into ChatGPT Plus and is particularly suitable for graphics with text, for example in the form of speech bubbles.
- Adobe Firefly generates very realistic images, but shows weaknesses when designing human faces.
- Stable Diffusion offers the highest customizability of all image models and is therefore often used by professionals.
This article was created in cooperation with Swisscom
Swisscom is committed to ecological, social and economic sustainability: climate protection, a sustainable lifestyle and the responsible use of digital media. Switzerland's leading ICT company has already received several awards for its long-standing commitment to sustainability and, according to "TIME Magazine", is one of the 500 most sustainable companies in the world in 2024. Swisscom Campus offers valuable knowledge and tips on digital media and their use in everyday life.
