“In the 21st century there are billions of pictures that look like undoctored photographs but are not — and viewers want to know what’s what when they encounter a remarkable picture.”

There are two ways of labeling images in response:

Labeling option #1:

Attach warning labels to all of the billions of AI/doctored pictures that look like undoctored photographs but are not.

The problem with option #1:

Anytime people see a remarkable picture without a label, they may assume it is undoctored and free of AIFI.

Labeling option #2:

Label pictures that actually are undoctored photographs.

The advantage of option #2:

Anytime people see a remarkable picture without a label, they may assume that it is doctored or has AIFI.