Automated copyright bots won't work

My latest Guardian column is online: "Why a rights robocop will never work." In it, I address the issue of automated copyright enforcement systems and why they are a bad idea:

It would have to perform with near-perfection: even if it turns out that it catches every single infringement except for video that is re-cut to 16:9 with letterboxing at 31fps, then all the pirates will just encode it that way and evade the filter, meaning that the system would generate an unacceptable level of false negatives.
In other words, all the money spent on the system would be for naught because it would fail to catch a significant proportion of pirates.
It would also have to be nearly perfect in regards to false positives - every time it misidentified a home movie of your kids' first steps or your gran's 85th birthday as Police Academy 29 or Star Wars: Episode 0, Jedi Teen Academy, your own right to use the Internet to communicate with your friends and family would be compromised - likewise unacceptable.

Link

I write books. My latest is a YA science fiction novel called Homeland (it's the sequel to Little Brother). More books: Rapture of the Nerds (a novel, with Charlie Stross); With a Little Help (short stories); and The Great Big Beautiful Tomorrow (novella and nonfic). I speak all over the place and I tweet and tumble, too.

More at Boing Boing

Eurovision 2013: An American in London

The technology that links taxonomy and Star Trek

Stephen

If YouTube complies with Viacom’s demand, they will loose Common Carrier status and be in a terrible position legally. Surely YouTube/Google is not THAT stupid. They would get sued and prosecuted by small minded people in small towns all over the US, and possibly all over the world.
johen

@Cory: “That’s what YouTube is promising and Viacom is demanding — if your video generates a false positive, it never sees the light of day.”

Ok, so? Why should YouTube have to allow every movie people want to upload? YouTube is providing their service for free. They would be within their rights (albeit silly looking) if they randomly deleted every 17th upload. People are simply not entitled to unfettered YouTube uploads.

The video sees the light of day, you just have to use something other than YouTube. Now if you are positing a ‘slippery slope’ argument, that may be a different matter, depending on the method one sends the video.
Dillo

I think this is a bit specious because we know from numerous extant examples already that the EntertainmentIndustrialComplex doesn’t give a damn about false positives.

They’re pretty confident that the “victim” doesn’t have the money to fight their lawyers so it’s not something they have to care about.
It’s all “collateral damage” in a war they wouldn’t have to fight if people just went to the cinema like they’re supposed to and didn’t go off and invent this danged IntardTubeWeb Thang.
Teresa Nielsen Hayden / Moderator

Stephen (13), the comment you’re talking about wasn’t suppressed. It was never posted, and isn’t recorded in the system. Church, Absent, I have to assume the same thing happened to both of you. My guess is that you ran into a glitch being generated by the attempt to launch the new version of Boing Boing.

Yeago, on the other hand, got disemvowelled.
Flying Squid

I’m one of the people who could be affected by this. I make The Skeletor Show which redubs old He-Man cartoons. Now Entertainment Rights, PLC, the owners of the footage, have had a lot of chances to figure out I’m doing it (apart from being on BoingBoing, I’ve been in hundreds of newspapers, radio, etc.) so either they have their heads in the sand or, more likely, consider what I’m doing fair use since I don’t make a dime off of it, I just do it for fun, and it doesn’t hurt their product. I have even tried to contact them about getting the rights so I can sell it and they didn’t respond. I take that as a sign that they don’t really care about what I’m doing.

These sorts of measures will only upset fans. Entertainment Rights, PLC doesn’t seem to have a problem with me, but this new software would take my work off of YouTube.

Now I could be wrong. They could feel what I’m doing is piracy, but they have never sent YouTube a takedown notice or even suggested what I am doing is violating copyright.

It’s cases like mine where this sort of thing is going to essentially destroy what I do for my own (and others’) entertainment, something that makes no money and is just done in fun.

Imagine if they did this with Red vs. Blue. Initially they didn’t have any permission from Bungie Studies to use Halo footage, but it got so big and popular that Bungie started actively supporting them.

No one has ever had a problem with what I’ve done… well, not a legal problem, some people think it sucks, but that’s another story. I have used the show as an example of my work to send to prospective employers and have made a lot of friends from doing it. No one is losing money on what I do since they are seeing re-edited, re-animated, re-dubbed footage which, in the end, bears little resemblance to the original. This sort of software, if it worked, would block what I do.
jonathanpeterson

It’s always been obvious that automated copyright protection only works in a guilty until proven innocent model.

The only question is whether big media companies will ultimately succeed in lobbying for legislation of guilty until proven innocent models. So far they’re winning.
NoneSuch

There’s a couple of different angles you can take that would allow this system to work:

1) If it’s YouTube doing it, then you don’t have any “right” to post on YouTube and they are quite free and able to use a “guilty if we even suspect you might be guilty” model. They’re covering their own ass. If the system can remove 99% of copyrighted material and wipe out 5% of non-copyrighted material due to false positives, they’d probably find that acceptable.

2) The system could be intended to have a human componet, which would make sense. The program doesn’t remove any video but it does flag them as “possible copyright violations”. Some human comes along, looks at the video, looks at the source video that the program paired it with and decides if there was a violation or not. This system is designed to seperate the 5% of things that Need To Be Examined from the 95% that don’t.
Teresa Nielsen Hayden / Moderator

Squid:

Now Entertainment Rights, PLC, the owners of the footage, have had a lot of chances to figure out I’m doing it (apart from being on BoingBoing, I’ve been in hundreds of newspapers, radio, etc.) so either they have their heads in the sand or, more likely, consider what I’m doing fair use since I don’t make a dime off of it, I just do it for fun, and it doesn’t hurt their product.

Sorry to say, none of those circumstances make it fair use. They could still nail you any time.
Flying Squid

Well I’m not a lawyer, but as far as I know since I’m not making any money from it, nor ever intend do do so and don’t malign their product in any way, it’s fair use.
LJP

…except for video that is re-cut to 16:9 with letterboxing at 31fps, then all the pirates will just encode it that way and evade the filter…

If they do it right, the filter can be more sophisticated than that. Yes, pirates will be able to evade the filters through some combination of transformations, but it’s possible that the resulting video will be unwatchable and/or barely recognizable.

Would you want to watch a movie that had been inverted, time shifted, and played backwards?
Yeago

Thr’s lwys smthng Dctrw nlyss lvs t.

You’re overlooking the fact that, while it would generate false positives, these would all be submitted for human review and so, compared to the entire internet, the pile to sort through is much smaller.

But, dismissing this, you get a sexy, absolutist title.
abb3w

It’s not IMPOSSIBLE, it’s merely an AI Complete problem.

Presumably, a true Turing-test AI could qualify as a peer to humanity (though the courts would have a field day). So, any time there’s a question, assemble twelve AIs for an automated Jury, an automated judge, and automated lawyers for the Plaintiff and Defense (possibly identical save for the position they are serving). Presume one has been able to get the “court” Congressionally established, convinced the US President to nominate the “judge”, and gotten Senate approval (although “life tenure” on the federal bench becomes tricky when dealing with an AI). Voila, one ruling, which may stand as valid until appealed through conventional channels, until the law is changed, or until either the humans or AIs get fed up with the idiocy of modern society and revolt.

Not possible with the current state-of-the-art, certainly, but not inherently impossible.

Teresa Nielsen Hayden: Sorry to say, none of those circumstances make it fair use.

Err… sorta, but as I(AmNotALawyer) understand it, they certainly contribute strongly in that direction. That “it doesn’t hurt their product” would be “effect of the use on the market or potential market for the original work”; that Flying Squid “do[es]n’t make a dime off of it” indicates the “purpose and character of the use” is noncommercial, and possibly parody.

This does, however, leave the problem of “the nature of the work” (fictional works get a higher degree of protection, due to the higher level of creativity theoretically involved), and “the amount and substantiality of the portion used in relation to the work as a whole”. Redubbing the soundtrack completely (not using any of the original voices, theme, incidental music, etc.) would probably help, as would not using an entire episode for a remix. Nonetheless, ER PLC could doubtless sic lawyers on poor Flying Squid at any time.

Were I a juror in such a case, I’d look really unfriendly if ER PLC didn’t send a cease-and-desist before heading to court, and would like to hear what the lawyers have to say about their failure to respond to his attempts to initiate contact. But that’s just a question of collecting damages; even I can’t see that they wouldn’t have grounds (if they disprove fair use) to demand a injunction against the production or distribution of The Skeletor Show for the remaining duration of the He-Man copyright.

Teresa Nielsen Hayden / Moderator

Squid (5), Abb3w (8), “amount used in proportion to the work as a whole” was what I had in mind. It’s hard to establish fair use when you’re using an entire copyrighted work to produce a derivative work in the same medium.

When I said they could nail him any time, I didn’t specifically mean a lawsuit; just that they could take action on it. Squid’s not in the clear.

Church

w00t! One of mine made the cut!

Would you want to watch a movie that had been inverted, time shifted, and played backwards?

Sure, I’d just have to write a script to revert and reverse it (not sure where time-shifting comes in.) Not a bad idea, actually.

Flying Squid

Thank you, abb3w, for the long explanation.

Theresa: I’m not saying I am in the clear, that is actually the reverse of my point. ER PLC could easily send YouTube a takedown order or myself a cease and desist and I would abide by that without complaint.

The problem with the automated system is that they would get it taken off no matter what.

Here’s a great example, a Star Wars fan film. Lucasfilm acknowledges outright that fan films are not only acceptable as long as they are non-commercial, but in fact help their franchise. This automated system could knock those off of YouTube, thus actually hurting Lucasfilm.

Flying Squid

Pardon me. It’s Teresa. No H.

Nick

Does the Guardian not have an editor?
From the article: “Stamping out copyright infringement is harder than the old men in suits thinks it is.”
Followed immediately by: “It’s all the rage these days: crackpot proposals. . .”
Later, “. . .perform with near-perfection: even if it turns out. . .”
There may be more but I stopped looking. Sorry about the non sequitur but it’s weird to see errors like this in the (online) print of a popular newspaper. :P

Stephen

Apparently Boing Boing has the first automated filtering bot. It just disallowed my previous post on the grounds that it was “Wrong”.

Yeago

Yes Stephen, I got the same problem.

Another strange quirk is that the first line of my above comment “Thr’s lwys smthng Dctrw nlyss lvs t” has become “Thr’s lwys smthng Dctrw nlyss lvs t”

Funny, because it read fine just after posting. Oh, silly bots.

Church

I had the same “wrong” weirdness.

@Yeago: I thought that was a comment on the inability of bots to make sense of things that humans can!

But now that I think of it, wasn’t there a similar vowel-stripping glitch a few months back? Or was that a different site?

Absent

I think this could be the making of a new “game” of trying to make clips that cause false positives, just to mess with the system.

What the hell is the “entered the text wrong” gumph? I keep getting that message.

Cory Doctorow

Yeago@7: No, that’s not what’s being proposed at all — the idea is to have systems that *prevent you from uploading* the videos that infringe in the first place. That’s what YouTube is promising and Viacom is demanding — if your video generates a false positive, it never sees the light of day.

Yeago

@Church – I’m not that clever.