I'm inclined to agree that in a practical sense, there may be no foolproof way to prevent copying, but can I prove it? No, and I haven't heard any airtight proof yet.
Copying is inherent in normal computation, and it is irreversible. For example
X = A; // statement 1
X = B; // statement 2
When statement 2 is executed, there is no way to reverse it because X has no memory of its prior value. That is the essense of copying - forgetting that a copy was made.
From what little I know of quantum computing and cryptography, in that realm all processes are reversible, so it is possible to guarantee that copies can always be detected.
Back in the world of normal computation, if one can control the viewers of information, one can try to ensure that any copy is degraded and not as good as the original. For example, there is the watermark idea, which can be made practically invisible. Or additional information can be added that is not displayed, but which is required to show the image.
I'm not saying strong DRM is possible in normal computing. I'm just saying if it isn't, that's a strong claim, and I'd like to see an airtight proof of it. This field has a number of things once considered impossible, such as public-key cryptography and Dijkstra's mutex algorithm.