Except for Amazon's own imprints, "NOT Amazon" is selling you every ebook you buy from them. The ebooks are provided by the rights-holder who request DRM be applied or not.
There's no doubt that Amazon could probably be a little more proactive about weeding out the scammers, but then how far should anybody be reasonably expected to go in this regard? There is no Copyright Police force with a mandate to actively protect everyone's IP rights. The onus is still mostly on rights-holders themselves to protect their own property (through DMCA takedown notices and other legal means).
Aggressive/Proactive systems that have the ability to autonomously take down content are infinitely more scary to me than a few scammers making a quick buck here and there before they're shut down by a takedown notice or a cease & desist order. I remember a few years ago, such an automated 'bot took down the live-stream of the annual Hugo awards. The reason?? Copyrighted video of Dr. Who playing in the background when the show was being awarded its Hugo for the Best Dramatic Presentation (Short Form) category. Shut it down right before Neil Gaiman got up to speak, in fact (he had written the award-winning episode).
