A couple of weeks ago I discussed some news regarding an ex-prosecutor who was now testifying as an expert witness on behalf of those accussed of possessing child pornography (Prosecutors Threaten Child Porn Legal Defender). Federal prosecutors were very upset that this lawyer was disrupting their prosecutions. Under existing law, however, the defense the lawyer was providing (that the photos might be photoshopped and not actually of children at all) is perfectly legal and makes a lot of sense.

Later, Prof. Eugene Volokh wondered if the difficulties in prosecuting such cases would cause the Supreme Court to overturn the ruling that protects speech that may appear to be of children, but isn't (Child porn cases thrown out).

My argument has always been that the prosecutors just have to do a better job. As I noted before, they can create databases of authenticated child porn. You only need a handful of images for a successful prosecution, and if the target of the prosecution has a handful of images out of the authenticated database ... game over.

Additionally, the feds have to be a little more sophisticated about authenticating images. After all, digital manipulation of photos isn't limited to pornography alone. Luckily, it appears that other government departments have been taking some proactive steps.

Prof. Hany Farid, a researcher at Dartmouth College is developing algorithms that will help distinguish images that have been manipulated from images that have not (Investigating digital images):

Farid and his students have built a statistical model that captures the mathematical regularities inherent in natural images. Because these statistics fundamentally change when images are altered, the model can be used to detect digital tampering.

via Boing Boing

UPDATE
Bonus: Farid also works on "the digital reconstruction of Ancient Egyptian tombs."