I would echo your concerns.
I presume there's an issue with the international element that the Internet and deepfakes present. You could have a perpetrator who is in Russia or some small town in a country we don't have much familiarity with. It would be hard to potentially go after those perpetrators.
I did note in my paper that there is inherently a limitation in pursuing criminal prohibition of deepfakes. Again, the standard of beyond reasonable doubt perhaps would play into limiting who is convicted of these crimes, as well as the scope and the resources that would be required to actually provide and enforce criminal prohibitions of this kind of behaviour.
I think that with private law, civil remedies and things that are based on the balance of probabilities, potentially there is a wider scope for not criminal justice but justice of some kind and at least some kind of disincentivization for engaging in this kind of behaviour. That would be my answer.