Similar to text generators, image generators produce images based on a natural language description. Also, like text generators, these AIs are trained on massive amounts of data, in this case images that also have textual descriptions.
Due to the large statistical representation of non-diverse images used to train such models, such generators have had issues with the way they produce images of humans when using non-specific prompts. E.g., Asking for the picture of a CEO and the image generated is of a white man.
Software such as Adobe and Canva have built-in AI assistants that can generate images or graphics for your project.
Name |
What it does |
---|---|
DALL-E 3 (via ChatGPT Plus, Team, and Enterprise users) OpenAI |
Creates realistic and detailed images from text descriptions. |
Dream Studio (free) Stability AI |
Produces art, photographs, and illustrations from text prompts. Offers experimentation with different styles, smart filters, and editing features. |
MidJourney (subscription) | Can mimic art styles (art deco), artists, and genres (cartoons). Editing features include remix, inpainting, resolution. |
Meta AI (free) | Offers: basic styles (e.g. cartoon, realistic, crayon): moods (e.g. dreamy, mystical); lighting levels; and colour intensities. |
Craiyon (free) | Basic option categories are art (e.g. painting), drawing, or photo. Generates multiple images to choose from. |
Hemphill, T.A. (2024), "Editorial: Copyright protection, artistic imagery, and the adoption of responsible artificial intelligence principles", Journal of Ethics in Entrepreneurship and Technology, Vol. 4 No. 1, pp. 2-6. https://doi.org/10.1108/JEET-06-2024-025
Cunningham, J. (2025). Painting in gray: the legal and ethical ambiguities of AI-generated art. Journal of Information, Communication and Ethics in Society.