In isEazy Author you can use Artificial Intelligence to get complete pedagogical pieces based on your content with one-click technology: you can get your results without iterations in a conversation.
Our development team with the support of pedagogical experts has worked on preparing autoprompts for AI and defining several styles. You will only need to type in your prompt.
What do you need to know to write a prompt and generate your AI images in Author?
First, remember that the AI in charge of generating your images does not “understand” natural language. This means that it can process text and generate images based on it, but it does not have a deep understanding of the meaning, context, or human intentions behind the words. Put another way, the AI interprets your description as a set of patterns, without truly understanding the broader context surrounding those words or phrases.
To help you out, here are some dos and don'ts on how to write descriptions for AI image creation in isEazy Author.
Dos
1. Always write your prompts in English: You can use other languages, of course, but you will get better results in English.
Example in Spanish: “Un brazo robótico ensamblando piezas en una línea de producción mientras un ingeniero supervisa los ajustes en una tablet.”
Example in English: “A robotic arm assembling parts on a production line while an engineer supervises adjustments on a tablet.”
2. Be straightforward and specific: Clearly describe the main element of your image, the setting and the key elements. Notice the difference between these two examples:
“Operators working together in a factory.”
“Operators wearing hard hats and safety vests working together in a factory.”
3. Define relationships: If there are multiple characters, describe their interactions.
Example:
“A boy with a robot dog in a grass field.”
Example: “A boy playing with a robot dog in a grass field.”
4. Describe everyday scenes and objects: if you try to describe an element, scene or object specific to an area of knowledge, the AI will not know how to represent it. Try to re-conceptualize your idea with simpler elements.
Example: “A person who suffered a home incident.”
Example: “A person looks worriedly at the damp and mould on the wall of his house while making a phone call”.
5. Use descriptive language: Highlight physical characteristics or actions. AI will give more prominence to the elements to which we attribute a quality. If relevant, describe the atmosphere or emotional tone, powerful and evocative words are capable of conveying a lot of meaning in a few syllables. Ancient, mystical or surreal, for example, can significantly alter the interpretation and outcome of your description.
Example: “A dragon perched on a mountain top”.
Example: “A red, large-winged dragon, perched on a mountaintop, roaring fiercely.”
6. Use commas: separate the different elements of your prompt to make it clear and fluid.
Don’ts
1. Don't write excessively complex and detailed prompts: Try to synthesize, otherwise the result will not be up to what you expect, the sense of what is really central will be lost.
For example, instead of something like “Futuristic automotive factory in full production, with an automated assembly line manufacturing electric vehicles with aerodynamic design and shiny metallic finish. Yellow-black robotic arms assemble parts with precision, generating sparks as they weld. Engineers and technicians in corporate uniforms and safety glasses monitor the process on holographic screens, adjusting parameters and verifying quality in a high-tech industrial environment.”
You may enter “Futuristic automotive factory with robots assembling electric vehicles while engineers supervise.”
2. Don't use very abstract descriptions to represent concepts: Rephrase your idea and make it more concrete in your description with something more precise and tangible.
For example: instead of asking for a "Image representing proper stress management” …
You could write "A person sitting in meditation position on a rock in a natural landscape, with a soft sunset reflecting on a calm lake.”
3. Don't add negative descriptions: If you do this, the AI will really look at what you are denying and will end up capturing it in the images.
For example: “An underwater scene full of vibrant marine life, with colorful fish, coral reefs and sea creatures such as turtles, jellyfish and dolphins. The deep blue ocean is illuminated by sunlight filtering through the water. Do not add sharks”
4. Don't use words that refer to a style: Omit terms such as “painterly,” “realistic” or “cartoonish,” unless necessary. Focus on the concept itself. Use the styles we offer in Author.
5. Don't use words or expressions such as “represent” and “reminiscent of”: Keep in mind that the AI does not understand them.
6. Don't name specific places or specific geographic allusions. Example: “An aerial view of Paris”.
7. Don't ask for elements that contain numbers, codes, texts: some examples would be bills, coins, credit cards. If you need elements of these characteristics to appear in your image, consider choosing image styles where these types of details are not visible, such as Isometric or Flat Vector.
Here are some more sample prompts that you can try:
- A pile of books and a coffee cup on a cozy desk.
- Two friends hugging happily in the park, wearing casual clothes.
- An IT professional analyzing data on multiple screens.
- A company work team in a meeting at a modern office.
- Salesperson talking casually to a customer.
- A path leading to a mountain peak, representing career growth.