An argument against AI-rights

The topic of AI-rights is heavily discussed in films, debates and other media. Some conclude that ethically the right thing to do would be to ascribe rights to AI. Whether it is ethically right or not is not something I want to discuss here. My argument is directed at the practicality of AI rights, if one were to choose to do so.

If in the future a state were to ascribe rights to AIs, this would be accompanied by certain obligations for the state and its citizens to protect the rights of AIs (whatever the laws would be exactly). But then someone could come along and load 1000 AI instances onto a 1 exabyte hard drive. The person could blackmail the state into giving him money or he would destroy the hard drive. Even if the person were imprisoned, he could keep up the threat using a dead-man-switch and thereby cheat his way out of jail. Thus, it is shown that a practical implementation of AI rights is not possible.

What do you think of the argument or what speaks against it? I can't think of any other way, but you might. I would like to hear your ideas