Developers Forum for XinFin XDC Network

Cover image for [Informative] Using XDC network to solve AI generated image tracking
Code and Music
Code and Music

Posted on

[Informative] Using XDC network to solve AI generated image tracking

One of the biggest challenges with AI that creates images is figuring out whether an image was made by a computer or a person. To solve this, we need a strong system that can verify if an image was AI-generated.

Many people are looking into ways to mark these images, like adding a digital watermark, but I think blockchain technology could offer a better solution.

Here's an idea for keeping track of AI-generated images:

Modify and launch a new diffusion model that will create a hash of the image and save it in the xdc network before the image is converted from binary/raw to jpeg/png. This is the last layer to be added to the model to make sure that before an image is generated the image hash is saved in a smart contract on the XDC network.

To validate an image: With this setup, anyone could use a tool called the XDC Image Explorer to check if an image was created by AI. You'd just upload the image, and the tool would look for its unique code on the blockchain to confirm whether or not it was AI-generated.

How to do it:

Here is an architecture on how it will work
Image description

This is on follow up on @riteshkakkad linkedin post.

Happy to answer any further questions.

Discussion (0)