Using AI text-to-image to bring student work to life

My daughter is in second grade and loves to write and draw. I created several AI generated ebooks for her when I was playing with text to image AI. You can read them by clicking the links below.

For the newest book, I took my daughter's story and created a book. The previous books were AI generated, however, this new book my daughter wrote herself.

At the end of the book, you can see her original story and pictures she drew. She underlined the rhyming words in her story so I kept it in the book.

I used to create the images for this new book. I took a different approach from previous books than I this book, I kept the main part of the prompt and just added or deleted parts of the prompt to get a more consistent look throughout the images. I also use the "Remix" option for each image. After an image is generated on, you can click on the three dots (...) to get more options. I selected the "remix option" each time to get a slightly different image for each page.

I used rather than other text-to-image generators such as Adobe Express or Microsoft designer due to the quality of the output desired to create a storybook image of a "pandacorn". As in all AI, it needs to be "trained", the concept of a pandacorn was not received well by the previous mentioned generators.

I think with all AI, trying different AI engines will give different results. One just needs to find results that work for you.

Have fun playing and feel free to reach out if you have any questions.

2 replies

March 29, 2024

Congratulations to Meghan! Tell her she is a very accomplished author - loved the story and great idea to use AI for the images. It’s nice to have a book artist in one’s pocket. And Meghan did a great job with her art too!

March 30, 2024

Great job Michael!

I love your creativity and you have created something that your daughter will love to read and look back on when she gets older.

Yes. The GenAI tool you use definitely makes a difference. The bigger data set the model was trained on, the better the image generated. Prompting with the correct word makes a great result also. You'll likely find that Adobe Firefly will net the best results.

Thanks for sharing!


Maximum file size: 400MB

Insert a video

Insert an image

Insert an image

125: 125
220: 220

This action can’t be undone.

Error Message

Are you sure you want to continue? Your changes will not be saved.

This post contains content from YouTube.

If you choose to view this content, YouTube may collect and process certain personal data. You can view YouTube’s <a href="" target="_blank">privacy policy here<span class="a11y">(opens in new window)</span>.</a>

This post contains content from YouTube.

You have rejected content from YouTube. If you want to change your consent, press the button below.