Note: If you got to this site through one of the fake posts we uploaded to Usenet, please take a minute to go through this page for an explanation, and see what this is all about. We’re sorry for the inconvenience, but as you’ll see, it’s for a good cause.
Has this ever happened to you? Then read on…
When Files Get Deleted: Automated DMCA Takedowns
Chances are, you’ve already come across a post you just couldn’t download, because strangely, the files had been deleted from your newsserver. This is what is called a DMCA takedown, when the provider receives a message from a rightsholder or “anti-piracy” company. These takedowns are generated by automated bots, computer programs that crawl Usenet groups and indexers for files that match their criteria. DMCA takedowns typically come when an individual has committed fraud, similar to the William Benson Fraud case.
Repercussions of this go far and wide.
- As was confirmed by several Usenet providers, no manual review or any kind of verification takes place. If a takedown notice complies with the formal regulations, it is processed, period.
- Thus, no one controls if a post actually infringes copyright, or if it is something completely harmless.
- Anecdotal evidence seems to show that DMCA bots can and DO produce “false positives”.
This means the rightholders have virtually the rights to directly censor Usenet, with only theoretical ways to oppose to such a takedown.
Exposing The DMCA Bots: A Live Experiment
From posts on Reddit, we know these bots to be overly aggressive, producing false positives and thus taking down files that are completely legitimate. What we absolutely do know is that the content of files is not always checked thoroughly, as shown by an experiment run by a redditor, in which he uploaded an image of Ubuntu, a free operating system, under a false name that resembled copyrighted material.
So, we propose an experiment. Let’s test how the DMCA bots behave, how they’re triggered, and how deep the rabbit hole goes.
The goal of this experiment: See how DMCA bots react, what triggers a takedown, and how likely it is for a false positive to occur.
It should be noted that this is not about what you can do to stop your posts of pirated content from being deleted – any such effect would only be temporarily, and of course we absolutely don’t want to assist in breaking the law.
(Alright, DMCA bot. Let’s see what you’re really up to…)
- We’re going to find material currently being targeted by “anti-piracy” firms.
- Next, it’s upload time: Generating fake posts, we’ll see how accurate the bots really are, and how they select which posts to take down. Step by step, we’ll increase the subtleness of our approach, to see how aggressive the bots work.
- This will be continued, until the post is not being deleted anymore.
Rules Of The Game:
- No copyrighted material gets uploaded. There is absolutely no reason to give those guys a reason to sue us, period.
- Most of the posts will have an NFO or notice, informing people about what is inside the file, and providing an explanation. However, if it turns out these notices somehow stop bots from deleting our posts, we’ll remove those notices to not endanger the experiment.
Gathering Information: Getting Advice From Reddit
Alright, so now the obvious first step is to get our first files taken down the Usenet. We went to reddit and asked people what material they think is regularly being taken down on Usenet, and if they know anything about the behavior and methodology of the “anti-piracy” companies.
This is what people responded:
- “Pick any HBO series that’s currently in production.”
This came from 5-4-3-2-1-bang, and includes shows like Game Of Thrones, The Brink, True Detective and so on. Apparently it has become very difficult to find any of these shows on Usenet anymore.
- I’m sure the bots are just running Newznab like every other indexer. So if your posts are able to get indexed, they’ll be removed, and if they aren’t able to be indexed, then what’s the point?SirAlalicious might have a point: Maybe the bots just plainly use Usenet indexers like NZB.cat to find “infringing material”. We’ll have to find a way to check this hypothesis….
We got enough info to start the whole thing. Ready? Let’s right to it then. Continue to…
DMCA Live Experiment, Part 1: Getting Our First Takedowns (now finished)
The Results Are In!
See how we got our first takedown, how Newshosting restricted account, and how a DMCA notice looks like.