A.I. - Artistic Integrity
Rebranding the definition of AI
Created on 18th January 2026
•
A.I. - Artistic Integrity
Rebranding the definition of AI
The problem A.I. - Artistic Integrity solves
AI-generated images are flooding online platforms, making it hard to tell what is genuinely human-made. This erodes trust, enables fraud, and exposes creators, platforms, and buyers to legal and reputational risk.
Our system solves this by verifying content authenticity. It uses cryptographic provenance (C2PA) when available to prove origin, and falls back to AI-based analysis when provenance is missing. Instead of blindly trusting editable metadata, it provides clear trust signals and confidence levels.
As a result, people can:
- Prove their work is genuinely human-made
- Safely commission and purchase art without fear of AI misuse
- Moderate platforms more effectively with less manual review
- Reduce legal and copyright risk
- Make informed decisions when authenticity matters
In short, it restores trust in visual content where metadata and guesswork are no longer enough.
Challenges we ran into
The main challenges that we ran into were the Machine Learning models that we tried to learn. Initially we had issues with the dataset being too small but with an addition of a 20k sized dataset a part of this issue was combatted.
We did have major issues with some models becoming overfitted having accuracies that remained at 50% or 99% and no where in between but after adjusting parameters and Epoch we eventually created a favourable ML
Technologies used