In the long run, if you don't want automated systems from downloading your stuff it would require a step where human intelligence has to be involved; any request that could be sent by a browser to initiate a download could be sent by a download accelerator. (And if said human intelligences were determined enough, then they could figure out what response to send and get their download accelerators to send that. Of course, such determination is no doubt comparatively rare.)
As drawmack suggests, you'll basically end up having to deal with the problem at your end - on the server - because you have no real control over what happens at the other. Continuing with the assumption that you're using an Apache server, a Google search on Apache bandwidth throttling turned up this among other things.
If you're using an ftp server then maybe that can be suitably configured also.
