Piracy tracking firm MUSO recently asked Google to remove tens of thousands of non-existing URLs, including those of mainstream sites such as Amazon, Variety, Billboard, the Huffington Post, and ABC News. The company explains that the incorrect DMCA notices are the result of a misconfiguration that has since been addressed. Meanwhile, the company is working with Google to ensure that the incorrect reports are retracted.

red crossThe UK piracy tracking firm MUSO has quite a balanced view on the piracy problem, framing it as an opportunity rather than a threat.

MUSO believes that by understanding what motivates pirates, copyright holders can gain great insights that will ultimately help to generate new revenue streams.

At the same time, however, the company offers classic anti-piracy services. This includes sending DMCA notices to Google for potentially infringing links that appear in the popular search engine.

Working on behalf of prominent rightsholders, MUSO has reported more than 100 million infringing URLs to Google over the years. This usually goes quite smoothly but this week we were alerted to a concerning pattern.

Non-existing URLs
In a series of quite unusual DMCA notices the company reported more than 100,000 URLs that flagged non-existing links on pirate sites. These URLs pointed to hundreds of domains and often used repetitive keyword strings, such as the examples shown below.

When we attempted to load these pages they nearly all resulted in 404-type errors. In some cases that’s not really a surprise, including the questionable example below.

questionable
These takedowns don’t really harm the targeted sites since the content doesn’t exist. However, more flagged links on a domain could be that these sites are further demoted in Google’s search results. That’s a serious concern.

Takedowns Target Legitimate sites
The issue is not limited to pirate sites either. Tucked away in the long lists of ‘pirate’ links we also find several legitimate websites. These include ABC News, Amazon, Billboard, Redfin, Variety, the Red Cross, and the UK National Archives.

Below are a few examples from a single notice, but there are many more.

http://amazon.co.uk/3253571-after-we...x264-xbet.html
http://www.abcnews.go.com/search-mov...ns%202021.html
http://billboard.com/3253571-after-w...x264-xbet.html
http://huffingtonpost.com/3253571-af...x264-xbet.html
http://globalnews.ca/3253571-after-w...x264-xbet.html
http://redcrossblood.org/3253571-aft...x264-xbet.html
http://redfin.com/3253571-after-we-f...x264-xbet.html
http://discovery.nationalarchives.go...x264-xbet.html

Misconfiguration
We reported our findings to MUSO earlier this week and the company informed us that these URLs should not have been targeted. Apparently, a batch of incorrect listings were sent out as the result of a misconfiguration

“These delistings are part of a batch of incorrect delistings sent out due to a misconfiguration within our service. We have identified the incorrect delistings and issued retractions with Google to have these incorrect delisting requests voided,” a spokesperson said.

Following up on this answer we requested a full overview of the incorrect delistings. In addition, we asked whether MUSO checks whether a URL is responding properly before a notice is sent.

While we didn’t get a direct answer to these questions, MUSO said that there are multiple processes and checks in place to identify mistakes. The company also reaffirmed that all mistakes will be retracted.

TorrentFreak spoke to the operator of one of the affected sites. He didn’t immediately notice a negative effect from the takedown notices. However, he points out that Google’s transparency report sometimes suggests that the URLs are indexed by the search engine, which isn’t the case.

Prevous Issues
From what we can gather, the takedown system wrongfully matched keyword strings to the wrong base URLs. This can work for some proxy sites that use the same URL structure but, in this case, something clearly went wrong.

muso takedown
This reminds us of a similar issue that popped up a few years ago. At the time, keyword string takedowns targeted many non-existing URLs as well but legitimate sites were not hit at the time.

While these types of mistakes should be avoided, we have to applaud MUSO for being open and transparent about it. Mistakes can happen and it’s always better to focus on a fix rather than a cover-up.