Twitter sued for $250M over copyright infringement

Kisai · June 18, 2023

20 hours ago, Spotty said:

Why must it be done automatically? 300,000 infringement notices over a period of 18 months is 550 notices per day. Those notices came from 17 publishers. An average of 32 claims per day per publisher. Some publishers would have a larger library of music and will be responsible for more notices than some others, but a total of 550 notices per day on average spread across the 17 publishers is entirely within limits of what can be done with manual review. Keep in mind these major publishers have entire departments of staff to deal with protecting their IP (or outsource to third party companies that deal with it).

Speaking from someone who has worked for a company that receives DMCA's, they are often FAX'd because the legal process demands a hard copy. I don't know if things have improved since then, but generally there would be a fax sent, and then an OCR that scans the document. So if the recording industry's representative did the most obtuse thing and sent 300,000 faxes (which they absolutely would do, and seeing as how they printed material for this lawsuit, likely did exactly in the form as seen)

The "automatic" process is for companies that have a specific reputation for following the rules and not shot-gunning everything. So generally these arrive as emails and a list of URL's, and the manual step in the middle is the person who receives it reviews every single item. In practice if the DMCA letter fills all the necessary legal statements, they will just take everything down, without even looking at it, and kick the responsibility to the user to fight back. The company receiving the DMCA is not obligated to push back and loses safe-harbor if they even "review" it.

So again, I sincerely doubt that all "300,000" notices fit the criteria, and their lack of sharing even one of these notices is suspect to me. This looks like a shakedown.

Now, did twitter maybe purposely ignore reports because their processes kinda suck? Yep, I'm pretty sure that's the case. Many companies "DMCA" process is one person who is an actual lawyer, and 10 people (for a company the size of twitter) for the ENTIRE WORLD actually doing the takedowns. It only needs to meet the criteria for a takedown in the US, nothing more. For countries that do not speak english, there typically needs to be someone who natively speaks that language to review the actual document to meet the criteria in english. But it's still getting pushed to someone who has access to remove or disable the data.

With that said, the statistical average of "a DMCA claim is valid" tends to be around 80%. Most claims are valid without needing to argue fair use, because entire images, videos, etc are uploaded unaltered. It's that other 20% that need to be reviewed, because there will either be an inconsequential use of the audio, or the takedown is in bad faith (Eg not the real copyright holder (eg competition,) or someone simply doesn't want the image, text, video or audio to exist because it embarrasses them (such as things the previous US president wanted taken down), or the material is taken out of context to "cancel" someone.)

Most MEME's start as copyright infringement before someone makes it look like garbage and posts it on the internet. If the company is expected to just take down every single claim, then anyone doing "meme" or "image macro" or "reaction gif" would have their accounts killed pretty darn quick. Because that happens. I've seen people I follow get suspended for doing some rather tame stuff, but it gets reported as "violent threats" and SOMEHOW that earns them a suspension when they are just MEME'ing on themselves or with friends who know it's a joke.

It can't be understated how much "human oversight" is required to read reports in the context they are posted in, and some companies, make the "DMCA" process extremely difficult in order to frustrate copyright holders because they know their business is enabling copyright infringement (eg cloudflare, imgur, tumblr, reddit, and twitter) by leaning on "we are only a proxy" or "we aren't responsible for what users upload" excuses.