Think together · Start with a story
A girl describes “a robot planting a tree in a garden,” and a model begins to create a new image. How does it turn that description into pixels?
Break down the process
- Text promptDescribe the scene
- Model generationBuild an image representation step by step
- PictureProduce a possible result
- Human reviewCheck details and requirements
Image-generation models use patterns learned during training to create pictures from prompts.
Common diffusion methods gradually remove noise from a noisy image or related representation; other generation approaches exist too.
The model does not necessarily pick an original picture from a gallery and paste it in. But it may create something similar to existing work, so review the result and use it responsibly.
Your turn to try
Could an image of “a robot landing on the Moon” be used as news evidence?
Ready? Open to see the explanation
No. It is a creative output and should be labeled as generated, not passed off as a real record.
One idea to remember
Image-generation models use patterns learned during training to create pictures from prompts.
Try covering the picture above and explaining each step to your family. If one step is hard to explain, revisit that card.