How to know which prompt was used to generate an AI image

In this post we are going to try to do some amateur reverse engineering . Imagine that you see one of the many great images on the internet right now, and that it has been generated by Artificial Intelligence with a tool like DALL-E, Midjourney or Stable Diffusion . Wouldn’t you like to be able to create a similar image yourself?

How to know which prompt was used to generate an AI image

The normal thing is that if it is a really elaborate illustration or digital image, the author has not shared the prompt or description they have used. In fact, this is quite a business and there are websites like PromptBase where people sell their prompts for prices that can be between 3 and 5 dollars.

How to obtain the detailed prompt that describes an image generated by Artificial Intelligence

Calm down, prompt sellers! I’m not going to spoil your party. As far as I know, there is no way to extract the exact prompt from an image unless its author has made it public.

Now, if we have used an AI to generate an exceptional image, we can also use another AI to do the reverse process . For this example we are going to use Copilot, since thanks to one of its latest updates it is now also capable of recognizing images.

Of course, we could also use ChatGPT Pro’s image recognition, but that’s a service that’s only available in the paid version, so we’ll opt to use Microsoft’s AI.

• Open the Copilot app. You can use the web version , the Copilot for Edge button, or the Copilot mobile app.
• Click on the camera icon that appears inside the text box to write. Select the AI ​​image you want to analyze.
• Now ask Copilot to return a prompt that is as detailed as possible to describe that image. You can use the following request, for example: “ Write a precise and detailed prompt that serves to generate an image as similar as possible to the image I just showed you .” You can also ask them to take into account other details such as perspective, camera focus, or artistic style.
• After a few seconds, Copilot’s AI will return an elaborate prompt that describes the image as faithfully as possible.

The Role of Prompts in AI Image Generation

A prompt in AI image generation refers to the input provided to the AI model, which guides the image creation process. It can take various forms, such as a specific phrase, a set of keywords, or even an existing image. The prompt defines the desired characteristics, style, and content of the generated image. By modifying or tweaking the prompt, an AI artist can influence the outcome and tailor it to their desired result.

How to Identify the Prompt Used in an AI Image

While determining the exact prompt used in generating an AI image is a complex task, there are a few techniques you can employ to gain some insights:

1. Reverse Image Search: Begin by performing a reverse image search using the AI-generated image as the input. This search might provide you with clues regarding the prompt used, especially if similar images or their sources appear in the results.
2. Analysis of Image Content: Carefully analyze the content of the image itself. Look for any distinctive features, objects, or themes that could indicate the initial prompt. For example, if the AI-generated image depicts a blue sky with fluffy white clouds, it is possible that the prompt contained keywords related to a clear day or pleasant weather.
3. User-Provided Information: If the AI-generated image was produced using a platform or a website, check if the user provided any details or descriptions regarding the prompt. Often, the artist or creator might offer insights into the inspiration or keywords they used to generate the image.