[jprete]

Data collection for adversarial training was also my first thought. The same training data used for classifying images as AI-generated can also be used to train the AI to generate images that fool more people.

[htrp]

Well the company's job is to detect deep fakes/photoshops. From their site.

Detecting Authenticity in the age of AI We detect AI-generated images to protect the integrity and authenticity of your service.

[john_noumenonic]

True enough, though we're interested in stress testing our model to see where the gaps are right now.