Guilty until proven innocent
September 3, 2012 3:23 PM Subscribe
How copyright enforcement robots killed the hugo awards Ustream's automated copyright takedown bots killed the legitimate coverage from the Hugo Awards ceremony.
Jesus Christ, what an artificial asshole.
posted by bz at 3:27 PM on September 3, 2012 [2 favorites]
posted by bz at 3:27 PM on September 3, 2012 [2 favorites]
I have a friend doing IT for the next WorldCon. He's already vowed to host the video on WorldCon controlled servers.
Not sure if that will be better or worse, since it might not stay up due to bandwidth, but it won't be shut down by a Video Provider's bots.
posted by Mad_Carew at 3:31 PM on September 3, 2012 [6 favorites]
Not sure if that will be better or worse, since it might not stay up due to bandwidth, but it won't be shut down by a Video Provider's bots.
posted by Mad_Carew at 3:31 PM on September 3, 2012 [6 favorites]
Robots killed the scifi convention broadcast? Looks like sci if has become obsolete.
posted by blue_beetle at 3:36 PM on September 3, 2012 [19 favorites]
posted by blue_beetle at 3:36 PM on September 3, 2012 [19 favorites]
It's hilarious that we've had the idea of broadcasting for quite awhile and the internet for long enough (if you ask me) but we still struggle and trip over ourselves trying to put it all together.
It's also stuff like this that makes me deeply distrustful of relying on technology for too much. It's still programmed by people and people don't do things right.
posted by bleep at 3:37 PM on September 3, 2012
It's also stuff like this that makes me deeply distrustful of relying on technology for too much. It's still programmed by people and people don't do things right.
posted by bleep at 3:37 PM on September 3, 2012
Yes, but Neil Gaiman wasn't getting paid for them showing clips! Which, unless he had a unique 'residuals for everything' contract, he wouldn't have been anyway. I suspect the BBC Doctor Who shows weren't even in the CopyrightBot's database... it was the clip from the nominated Community episode that triggered it.
posted by oneswellfoop at 3:37 PM on September 3, 2012
posted by oneswellfoop at 3:37 PM on September 3, 2012
Feel free to rail at the bots and Ustream all you like, but the person who really needs to know about your discontentment with the current implementation of copyright law is your U.S. congresscritter. I promise you they'd rather spend engineer-hours on something that makes them money, but the price of doing business in today's regulatory environment is to be nonsensically proactive about removing content that you suspect may be infringing.
posted by town of cats at 3:42 PM on September 3, 2012 [14 favorites]
posted by town of cats at 3:42 PM on September 3, 2012 [14 favorites]
We still don't really know what happened here, do we?
posted by mediareport at 3:53 PM on September 3, 2012
posted by mediareport at 3:53 PM on September 3, 2012
In the comments on io9 there's a lot of comments to the effect of "People, listen, the problem isn't ___, the problem is ____!" Actually, people, listen, there are a lot of problems with this situation.
A. The fact that this "Shoot first, ask questions never" approach is required by the current copyright climate and ustream felt compelled to spend time complying
B. The fact that there was no recourse to correct the problem quickly
C. The fact that there doesn't have to be any recourse at all
D. The no-nothing child-s-school-report approach to blog-journalism. What was the relationship between the Hugo Awards and Ustream? Did they have a contract or something? Who was in charge of this?
posted by bleep at 3:54 PM on September 3, 2012 [9 favorites]
A. The fact that this "Shoot first, ask questions never" approach is required by the current copyright climate and ustream felt compelled to spend time complying
B. The fact that there was no recourse to correct the problem quickly
C. The fact that there doesn't have to be any recourse at all
D. The no-nothing child-s-school-report approach to blog-journalism. What was the relationship between the Hugo Awards and Ustream? Did they have a contract or something? Who was in charge of this?
posted by bleep at 3:54 PM on September 3, 2012 [9 favorites]
They also killed the radio star.
posted by Nomyte at 3:55 PM on September 3, 2012 [2 favorites]
posted by Nomyte at 3:55 PM on September 3, 2012 [2 favorites]
Huh. I watched it all a few minutes later in a torrent from thepiratebay.org
posted by fredludd at 4:09 PM on September 3, 2012 [9 favorites]
posted by fredludd at 4:09 PM on September 3, 2012 [9 favorites]
Copyright is so fucked up. It's going to stay that way. Things like this will happen more and more often. *Shrugo.*
posted by cjorgensen at 4:16 PM on September 3, 2012 [1 favorite]
posted by cjorgensen at 4:16 PM on September 3, 2012 [1 favorite]
Bleep -- bloggers don't have time for your so-called "sources" and "investigative journalism". There's a story to tell and they're gonna tell it RIGHT NOW. I mean, they always post clarifications and retractions, right? RIGHT?
posted by bpm140 at 4:30 PM on September 3, 2012
posted by bpm140 at 4:30 PM on September 3, 2012
With gawkermedia in particular lately I feel as if I'm going to be reading some high schooler's homework i should be getting paid for it.
posted by bleep at 4:33 PM on September 3, 2012 [3 favorites]
posted by bleep at 4:33 PM on September 3, 2012 [3 favorites]
The future is a world of badly programmed bots.
posted by jaduncan at 4:33 PM on September 3, 2012 [1 favorite]
posted by jaduncan at 4:33 PM on September 3, 2012 [1 favorite]
How is IO9 so sure that bots cut off the stream and not an actual human making a bad assumption?
posted by Ardiril at 4:40 PM on September 3, 2012
posted by Ardiril at 4:40 PM on September 3, 2012
I, for one, welcome our new robot IP enforcement overlords!
posted by monotreme at 4:41 PM on September 3, 2012 [1 favorite]
posted by monotreme at 4:41 PM on September 3, 2012 [1 favorite]
For a moment I felt so powerful. I was living in the future! Nerd TV was streaming straight onto my screen! There from the comfort of my own living room I was eating trail mix, drinking a gin and tonic, and enjoying the verbal stylings of Metafilter's own John Scalzi.
Then Neil Himself came on stage, and they tore it all away. I was left frantically hitting refresh on Twitter, sobbing into my almonds and cursing U-stream.
The future sucks
posted by crackingdes at 4:47 PM on September 3, 2012
Then Neil Himself came on stage, and they tore it all away. I was left frantically hitting refresh on Twitter, sobbing into my almonds and cursing U-stream.
The future sucks
posted by crackingdes at 4:47 PM on September 3, 2012
Does Ustream run bots supplied by major companies without knowing how well those bots work? Do they know, and just not care about the edge cases?
Or do big content owners run bots on the internet, send microsecond cease-and-desists that are then (poorly) evaluated by microsecond robot lawyers who then shut off the feeds?
I'm really curious how this works, and that link doesn't really give any information.
posted by nathan v at 5:00 PM on September 3, 2012 [1 favorite]
Or do big content owners run bots on the internet, send microsecond cease-and-desists that are then (poorly) evaluated by microsecond robot lawyers who then shut off the feeds?
I'm really curious how this works, and that link doesn't really give any information.
posted by nathan v at 5:00 PM on September 3, 2012 [1 favorite]
I, for one, welcome our new robot IP enforcement overlords!
It's Brazil with bots.
posted by jaduncan at 5:14 PM on September 3, 2012 [2 favorites]
It's Brazil with bots.
posted by jaduncan at 5:14 PM on September 3, 2012 [2 favorites]
Would have been tricky, but not impossible, on a live event -- but I've heard that people change the speed and pitch of videos by 1% to avoid matching the copyright templates.
posted by StickyCarpet at 5:32 PM on September 3, 2012
posted by StickyCarpet at 5:32 PM on September 3, 2012
How is IO9 so sure that bots cut off the stream and not an actual human making a bad assumption?
If bots cut off the stream, I bet that some actual humans made bad assumptions in order to make that possible. Namely, the assumption that those bots would not do things like this as often as they do.
But, like Nathan V, I would love to hear which humans those would be, or at least which company employs them, what costs they will entail as a result of this failure, and what benefits they had in making the assumptions that lead to this result.
posted by RobotHero at 5:33 PM on September 3, 2012
If bots cut off the stream, I bet that some actual humans made bad assumptions in order to make that possible. Namely, the assumption that those bots would not do things like this as often as they do.
But, like Nathan V, I would love to hear which humans those would be, or at least which company employs them, what costs they will entail as a result of this failure, and what benefits they had in making the assumptions that lead to this result.
posted by RobotHero at 5:33 PM on September 3, 2012
I'd love to know more about how it works too. The impression I get know is that I could create my own bot that randomly sent takedown notices and it would just work. Is nobody else tempted to start takedowning, say, Rebecca Black, or Will Smith's Wild Wild West (my own nominee for Worst Song Ever)?
posted by jacalata at 5:58 PM on September 3, 2012 [1 favorite]
posted by jacalata at 5:58 PM on September 3, 2012 [1 favorite]
Would have been tricky, but not impossible, on a live event -- but I've heard that people change the speed and pitch of videos by 1% to avoid matching the copyright templates.
One could (so I hear) also tilt the video by 1-2%. Visually imperceptible, and you can crop the first couple of lines of pixels on all sides to make up for it. If the bot is smart enough to catch that, you can accept the visual pain and tilt ~5%. Bonus point: it works fine on live streams.
posted by jaduncan at 6:00 PM on September 3, 2012 [1 favorite]
One could (so I hear) also tilt the video by 1-2%. Visually imperceptible, and you can crop the first couple of lines of pixels on all sides to make up for it. If the bot is smart enough to catch that, you can accept the visual pain and tilt ~5%. Bonus point: it works fine on live streams.
posted by jaduncan at 6:00 PM on September 3, 2012 [1 favorite]
Perhaps the bots did us favour in denying a live view of Gaiman getting awarded for that horrible episode from a horrible season.
But this is the sort of absurd event we'll probably see more of. If I was producing a comedy show I'd do a skit about buying a movie or music in 2020 or some such. Lawyers everywhere, a contract that is pages with signatures here and there and here again, an implant that also monitors your media consumption habits because it "speaks" to your media devices, and of course it would be an allegory on how drugs and media crime are filling our jails.
posted by juiceCake at 6:03 PM on September 3, 2012
But this is the sort of absurd event we'll probably see more of. If I was producing a comedy show I'd do a skit about buying a movie or music in 2020 or some such. Lawyers everywhere, a contract that is pages with signatures here and there and here again, an implant that also monitors your media consumption habits because it "speaks" to your media devices, and of course it would be an allegory on how drugs and media crime are filling our jails.
posted by juiceCake at 6:03 PM on September 3, 2012
Design, design, design.
If the software/API fails (and it appear it did here), then someone used a jackhammer to insert penny nails. 'Twas doomed to failure.
posted by pjmoy at 6:16 PM on September 3, 2012
If the software/API fails (and it appear it did here), then someone used a jackhammer to insert penny nails. 'Twas doomed to failure.
posted by pjmoy at 6:16 PM on September 3, 2012
Do the people programming the bots have any incentive whatsoever to reduce the number of mistaken positives? Isn't their incentive totally weighted to minimize (a) false negatives and (b) cost? From their point of view, there's no problem here.
posted by tyllwin at 6:20 PM on September 3, 2012 [1 favorite]
posted by tyllwin at 6:20 PM on September 3, 2012 [1 favorite]
"reduce the number of mistaken positives"
This isn't a case of a 'mistaken positive', rather this is a case of making the wrong the decision after identifying a positive. That is the reason I question whether pulling the plug was truly a bot behavior as opposed to a human reaction.
posted by Ardiril at 6:25 PM on September 3, 2012
This isn't a case of a 'mistaken positive', rather this is a case of making the wrong the decision after identifying a positive. That is the reason I question whether pulling the plug was truly a bot behavior as opposed to a human reaction.
posted by Ardiril at 6:25 PM on September 3, 2012
I meant false positive in a looser sense: incorrectly identifying an infringement. So, to reduce confusion, let me ask it this way: once they detect any tiny sample of copyrighted video, what earthly incentive do they have to care if it's fair use, or authorized, or not? Why would you pay for human to make the call? Why isn't your most profitable course to just always send a takedown?
posted by tyllwin at 6:32 PM on September 3, 2012 [1 favorite]
posted by tyllwin at 6:32 PM on September 3, 2012 [1 favorite]
It's a damned shame, because the rest of the Hugos were awesome. I spent half an hour sitting waiting for the escalator crush to clear reading twitter ranting and being sad for everyone who was not me.
(Not very sad. I was there! Wooo!)
posted by restless_nomad at 6:36 PM on September 3, 2012
(Not very sad. I was there! Wooo!)
posted by restless_nomad at 6:36 PM on September 3, 2012
This isn't a case of a 'mistaken positive', rather this is a case of making the wrong the decision after identifying a positive.
Still a mistaken positive, as they were not in violation of copyright law. It's no different than if it took down a video that was uploaded by the copyright holders themselves (that does happen too.) The problem is that it's impossible for a human to reliably make a fair use call, nevermind a bot.
That is the reason I question whether pulling the plug was truly a bot behavior as opposed to a human reaction.
Not likely. The sheer volume of positives precludes a timely (or any) review. Anything reacting that fast isn't likely human.
What we really need is for penalties for false takedowns to be as ruinous as actual violations. Right now there's no incentive not to act overcautiously.
posted by ChurchHatesTucker at 6:40 PM on September 3, 2012
Still a mistaken positive, as they were not in violation of copyright law. It's no different than if it took down a video that was uploaded by the copyright holders themselves (that does happen too.) The problem is that it's impossible for a human to reliably make a fair use call, nevermind a bot.
That is the reason I question whether pulling the plug was truly a bot behavior as opposed to a human reaction.
Not likely. The sheer volume of positives precludes a timely (or any) review. Anything reacting that fast isn't likely human.
What we really need is for penalties for false takedowns to be as ruinous as actual violations. Right now there's no incentive not to act overcautiously.
posted by ChurchHatesTucker at 6:40 PM on September 3, 2012
"what earthly incentive do they have to care if it's fair use, or authorized, or not?"
For contracted usage, the IP owner gets paid. You don't want a bot stopping that.
That is the problem with this slipshod reporting; basic questions like yours cannot be answered definitively because far too many details were omitted.
posted by Ardiril at 6:41 PM on September 3, 2012
For contracted usage, the IP owner gets paid. You don't want a bot stopping that.
That is the problem with this slipshod reporting; basic questions like yours cannot be answered definitively because far too many details were omitted.
posted by Ardiril at 6:41 PM on September 3, 2012
"Still a mistaken positive, as they were not in violation of copyright law."
It would be a mistaken positive if the bot thought the Doctor Who (or Community) clip was a clip from the Avengers (as an example). Identification and decision-making are two discrete actions.
posted by Ardiril at 6:46 PM on September 3, 2012
It would be a mistaken positive if the bot thought the Doctor Who (or Community) clip was a clip from the Avengers (as an example). Identification and decision-making are two discrete actions.
posted by Ardiril at 6:46 PM on September 3, 2012
It would be a mistaken positive if the bot thought the Doctor Who (or Community) clip was a clip from the Avengers
Well whether it's a false positive or not is a matter of what you're trying to identify. If you're trying to identify "any instance of Doctor Who video" then it's not a false positive. If you're trying to identify "any instance of copyright infringement," then it is. My point is that the incentives are heavily tilted in favor of simply treating "any instance of Doctor Who video" as "infringement," regardless of whether it actually is or not.
The downside of missing fair use is that the Hugos likely won't use this service again. The downside of missing infringement is potential loss of your business. I don't see how better reporting from Gawker changes or refutes that.
posted by tyllwin at 6:58 PM on September 3, 2012
Well whether it's a false positive or not is a matter of what you're trying to identify. If you're trying to identify "any instance of Doctor Who video" then it's not a false positive. If you're trying to identify "any instance of copyright infringement," then it is. My point is that the incentives are heavily tilted in favor of simply treating "any instance of Doctor Who video" as "infringement," regardless of whether it actually is or not.
The downside of missing fair use is that the Hugos likely won't use this service again. The downside of missing infringement is potential loss of your business. I don't see how better reporting from Gawker changes or refutes that.
posted by tyllwin at 6:58 PM on September 3, 2012
Identification and decision-making are two discrete actions.
They're supposed to be. That's even how the DMCA (supposedly) works. But, as in cases where Legal sues over something that was posted with the permission of Marketing, they're often not.
The problem is there are almost no incentives to get it right. Youtube installs software to police for content so valuable that the studios can't be bothered to do it themselves, and some mom gets flagged for filming her infant bopping to a Prince song.
No big deal to the studios, unless someone makes a stink about it. Pretty devastating to mom, though.
posted by ChurchHatesTucker at 7:11 PM on September 3, 2012 [1 favorite]
They're supposed to be. That's even how the DMCA (supposedly) works. But, as in cases where Legal sues over something that was posted with the permission of Marketing, they're often not.
The problem is there are almost no incentives to get it right. Youtube installs software to police for content so valuable that the studios can't be bothered to do it themselves, and some mom gets flagged for filming her infant bopping to a Prince song.
No big deal to the studios, unless someone makes a stink about it. Pretty devastating to mom, though.
posted by ChurchHatesTucker at 7:11 PM on September 3, 2012 [1 favorite]
From the article: "First of all, the clips had been provided by the studios to be shown during the award ceremony. The Hugo Awards had explicit permission to broadcast them. But even if they hadn't, it is absolutely fair use to broadcast clips of copyrighted material during an award ceremony."
Did the Hugo Awards indeed have explicit permission to broadcast those clips over UStream, or is this an assumption by the article writer? Is UStream limited to the specific territories, or is it global? Can we assume that the use of these clips was fair use throughout the territories UStream serves?
"If you're trying to identify 'any instance of copyright infringement' " - This is a common mindset among those who do not know the ways of programming artificial intelligence. Bots don't know shit about infringement; they only know how to measure a sample's signature and qualify that signature against a datafile to determine its fit. The strength of that fit is then compared to a decision table, and that decision table determines the subsequent action.
Better reporting would have provided us with some evidence that this was in fact a bot. As of now, that bots were involved is itself an assumption.
"They're supposed to be." - For a bot, they absolutely are. Once Legal gets involved, you are completely outside the cyber realm.
posted by Ardiril at 7:14 PM on September 3, 2012
Did the Hugo Awards indeed have explicit permission to broadcast those clips over UStream, or is this an assumption by the article writer? Is UStream limited to the specific territories, or is it global? Can we assume that the use of these clips was fair use throughout the territories UStream serves?
"If you're trying to identify 'any instance of copyright infringement' " - This is a common mindset among those who do not know the ways of programming artificial intelligence. Bots don't know shit about infringement; they only know how to measure a sample's signature and qualify that signature against a datafile to determine its fit. The strength of that fit is then compared to a decision table, and that decision table determines the subsequent action.
Better reporting would have provided us with some evidence that this was in fact a bot. As of now, that bots were involved is itself an assumption.
"They're supposed to be." - For a bot, they absolutely are. Once Legal gets involved, you are completely outside the cyber realm.
posted by Ardiril at 7:14 PM on September 3, 2012
"They're supposed to be." - For a bot, they absolutely are. Once Legal gets involved, you are completely outside the cyber realm.
Who's making assumptions now?
posted by ChurchHatesTucker at 7:19 PM on September 3, 2012
Who's making assumptions now?
posted by ChurchHatesTucker at 7:19 PM on September 3, 2012
I understand how bots work. Since the bot includes the decision logic, I disagree that they can't have some fair use intelligence built in. You could, for example measure percentages: what percentage of the suspect video matches the protected source? What percentage of the protected source is included in the suspect video. That's not sophisticated intelligence, but it's not a binary, and I don't see what difference it makes if that logic resides in the video matching routines or a later decision tree.
However, the question of whether or not it's a bot is entirely orthogonal to my concern. I'm perfectly content to assume that it was, or might have been a real human. My point is that whether it's a bot or a human, there's no reason for that entity to care whether it's infringement or not. Bot or human, the business incentive is always to treat any match as an infringement, unless you have a specific override.
posted by tyllwin at 7:35 PM on September 3, 2012 [1 favorite]
However, the question of whether or not it's a bot is entirely orthogonal to my concern. I'm perfectly content to assume that it was, or might have been a real human. My point is that whether it's a bot or a human, there's no reason for that entity to care whether it's infringement or not. Bot or human, the business incentive is always to treat any match as an infringement, unless you have a specific override.
posted by tyllwin at 7:35 PM on September 3, 2012 [1 favorite]
I disagree that they can't have some fair use intelligence built in.
Is there a definition/measure of fair use that's agreed to by most content owners, or that the law consistently recognizes and affirms? I thought it was a concept kept purposefully vague.
posted by Lentrohamsanin at 7:48 PM on September 3, 2012
Is there a definition/measure of fair use that's agreed to by most content owners, or that the law consistently recognizes and affirms? I thought it was a concept kept purposefully vague.
posted by Lentrohamsanin at 7:48 PM on September 3, 2012
The bad reporting aspect is that we don't know anything about the relationship between the Hugos and ustream. There wasn't even an attempt to get a comment out of either of them. We don't know if this was a good business decision on ustream's part or not.
posted by bleep at 7:49 PM on September 3, 2012
posted by bleep at 7:49 PM on September 3, 2012
UStream's explanation as to what happened, answering (at least in part) some of the questions asked.
posted by mephron at 8:04 PM on September 3, 2012 [4 favorites]
posted by mephron at 8:04 PM on September 3, 2012 [4 favorites]
This has nothing to do with copyright, and everything to do with stupid websites making stupid, cheap decisions. "Oh, it was the third party vendor who did it. [shrug] You get what you pay for, amiright?"
posted by gjc at 8:18 PM on September 3, 2012
posted by gjc at 8:18 PM on September 3, 2012
Thanks mephron. That note sound entirely reasonable, and it does answer a lot of questions.
posted by hippybear at 8:35 PM on September 3, 2012
posted by hippybear at 8:35 PM on September 3, 2012
I'm thinking that there's information missing based on some of the stuff mentioned in the comments (yes yes I know don't read the comments shut up), but I'd like to hear from some of the people who claimed they were told that the block wasn't going to go away.
That kind of situation would explain some things - webcomicer Jeph Jacques got his UStream taken down hard and he was told he was banned; if they did it because he was listening to music and that system picked it up and blocked it, that might be part of the reason why.
The big problem is, though, that these systems exist and that there is, apparently, no easy way to override them. That is something that really needs addressing.
posted by mephron at 8:54 PM on September 3, 2012 [1 favorite]
That kind of situation would explain some things - webcomicer Jeph Jacques got his UStream taken down hard and he was told he was banned; if they did it because he was listening to music and that system picked it up and blocked it, that might be part of the reason why.
The big problem is, though, that these systems exist and that there is, apparently, no easy way to override them. That is something that really needs addressing.
posted by mephron at 8:54 PM on September 3, 2012 [1 favorite]
yes yes I know don't read the comments shut up
Those comments are a little bit over-the-top, I think. Some excerpts:
Those comments are a little bit over-the-top, I think. Some excerpts:
In 6 months, you’ll be out of business, I think.And:
This is not going to just be forgotten; Ustream is going to be associated with “FAIL” in the fannish memory for a very long time. Perhaps the rest of my life. Perhaps even the rest of yours… And don’t think that science fiction fandom is just going to let this slide without telling anyone. Some of us can be strident, vindictive, and very persistent.posted by grouse at 8:58 PM on September 3, 2012
As mephron notes, the comments below the apology make a couple of interesting points:
1. Worldcon folks were told flatout by Ustream that the feed would not be started - "the people running the event were explicitly informed that Ustream had no intentions of lifting the ban" - which conflicts directly with the president's version that Ustream immediately "attempted to restore the broadcast."
2. The claim that it is not Ustream's responsibility to automatically take down possibly infringing content, but rather "the responsibility of author/owner of copy written material to file a DCMA claim asking that the infringing material be removed."
Yeah, there's some silly, over-the-top geek fury, but those two things seem worth pulling out and looking at. And then there's this:
Users of our paid, ad-free Pro Broadcasting service are automatically white listed to avoid situations like this and receive hands-on client support.
Hmm. So Ustream is making a conscious decision to not give users of its free service the legal benefit of the doubt - and is even going further than the law requires in automating takedowns to avoid having to devote any man-hours - in order to drive users to its paid service? Is that right?
posted by mediareport at 9:05 PM on September 3, 2012
1. Worldcon folks were told flatout by Ustream that the feed would not be started - "the people running the event were explicitly informed that Ustream had no intentions of lifting the ban" - which conflicts directly with the president's version that Ustream immediately "attempted to restore the broadcast."
2. The claim that it is not Ustream's responsibility to automatically take down possibly infringing content, but rather "the responsibility of author/owner of copy written material to file a DCMA claim asking that the infringing material be removed."
Yeah, there's some silly, over-the-top geek fury, but those two things seem worth pulling out and looking at. And then there's this:
Users of our paid, ad-free Pro Broadcasting service are automatically white listed to avoid situations like this and receive hands-on client support.
Hmm. So Ustream is making a conscious decision to not give users of its free service the legal benefit of the doubt - and is even going further than the law requires in automating takedowns to avoid having to devote any man-hours - in order to drive users to its paid service? Is that right?
posted by mediareport at 9:05 PM on September 3, 2012
(it should say "claims" and not "points" in that first line)
posted by mediareport at 9:05 PM on September 3, 2012
posted by mediareport at 9:05 PM on September 3, 2012
And don’t think that science fiction fandom is just going to let this slide without telling anyone. Some of us can be strident, vindictive, and very persistent.
"I, for example, am a dick."
posted by jaduncan at 9:09 PM on September 3, 2012 [5 favorites]
"I, for example, am a dick."
posted by jaduncan at 9:09 PM on September 3, 2012 [5 favorites]
Reading the comments on the Ustream blog post I am struck by the fury of the privileged. Living under a rock much? Wah wah wah - the stream got taken down. Welcome to the Internet! You just got chilled by effects.
posted by vicx at 9:34 PM on September 3, 2012
posted by vicx at 9:34 PM on September 3, 2012
From the UStream "explanation:"
I have suspended use of this third-party system until we are able to recalibrate the settings so that we can better balance the needs of broadcasters, viewers, and copyright holders.
How about you throw that shit out and make rightsholders go through the legally prescribed channel of filing a DMCA complaint? It's not like the process isn't lopsided enough.
posted by ChurchHatesTucker at 9:48 PM on September 3, 2012 [1 favorite]
I have suspended use of this third-party system until we are able to recalibrate the settings so that we can better balance the needs of broadcasters, viewers, and copyright holders.
How about you throw that shit out and make rightsholders go through the legally prescribed channel of filing a DMCA complaint? It's not like the process isn't lopsided enough.
posted by ChurchHatesTucker at 9:48 PM on September 3, 2012 [1 favorite]
"many video sites around the web"
I found a Vobile promotional video here:
http://www.youtube.com/watch?v=-qP7mRcINFE
It claims the same system is used by uStream, youku, todou, and justin.tv ( Those are all the ones it mentions by name, but you can check their logo garden around 3:45 for any other sites. )
posted by RobotHero at 9:52 PM on September 3, 2012
I found a Vobile promotional video here:
http://www.youtube.com/watch?v=-qP7mRcINFE
It claims the same system is used by uStream, youku, todou, and justin.tv ( Those are all the ones it mentions by name, but you can check their logo garden around 3:45 for any other sites. )
posted by RobotHero at 9:52 PM on September 3, 2012
How about you throw that shit out and make rightsholders go through the legally prescribed channel of filing a DMCA complaint?I'm assuming this is because the various XXAAs have threatened to sue them into the ground if they don't
posted by fullerine at 10:08 PM on September 3, 2012
This is where an end-to-end encryption plugin might work very nicely. Some kind of object wrapper the unencrypts the video feed, but then again, how to you prevent the bots from downloading the public key and unencrypting..Maybe a captcha of some sort?
If you fail the captcha test, all you'd see is video noise...
Yeah, the PKI backend could be hosted by a completely separate system than the video store.
posted by roboton666 at 10:46 PM on September 3, 2012 [1 favorite]
If you fail the captcha test, all you'd see is video noise...
Yeah, the PKI backend could be hosted by a completely separate system than the video store.
posted by roboton666 at 10:46 PM on September 3, 2012 [1 favorite]
>How about you throw that shit out and make rightsholders go through the legally prescribed channel of filing a DMCA complaint?
I'm assuming this is because the various XXAAs have threatened to sue them into the ground if they don't
Someone's got to realize eventually that pushing back is a market difference.
posted by ChurchHatesTucker at 11:04 PM on September 3, 2012
I'm assuming this is because the various XXAAs have threatened to sue them into the ground if they don't
Someone's got to realize eventually that pushing back is a market difference.
posted by ChurchHatesTucker at 11:04 PM on September 3, 2012
It's only a matter of time before a ustream equivalent uses a third party system that's hooked into a drone aircraft armed with missiles. "We are sorry the 2020 Hugo awards ceremony was hit by hellfire missiles. We are suspending our operations with IPOViolatorKiller until the algorithms can be adjusted."
posted by happyroach at 11:53 PM on September 3, 2012 [1 favorite]
posted by happyroach at 11:53 PM on September 3, 2012 [1 favorite]
Clearly it was the cybermen
I still blame Oswin.
posted by homunculus at 12:04 AM on September 4, 2012
I still blame Oswin.
posted by homunculus at 12:04 AM on September 4, 2012
Okay, so third party. Ustream hires Vostream, presumably to reduce workload (but probably also as part of some contracts with major content providers). Vostream is a little lazy, doesn't do any "fair use" checks (which I would imagine would be best done in software by looking at for how long the clip in question is played). Since Ustream doesn't legally have to react in microseconds to a cease-and-desist, and they're just using Vostream for reducing labor costs, part of their premium service includes human review rather than bot review, which Ustream probably thought wasn't even useful, except as a delay to take-down-- mostly just a bullet point without any presumed value.
Is that an accurate takeaway?
So this is a great thing that happened. Vostream loses a bit of business, adds a few fair use caveats to its bots. Ustream changes the third party it chooses to license software from, presumably to one with fair-use caveats.
Interesting, still, to learn about companies like Vostream. That's got to be complicated work, and it's not even required by current law (as far as I know, I'm not a lawyer). I bet the business model is as much about charging for registration of copyrighted material as it is about charging for copyright violation detection. Now that I think about it, I bet Ustream doesn't actually pay anything for Vostream, and Vostream collects payments from content providers based on how many internet delivery sites agree to use it. Kind of a clever business plan.
posted by nathan v at 1:46 AM on September 4, 2012
Is that an accurate takeaway?
So this is a great thing that happened. Vostream loses a bit of business, adds a few fair use caveats to its bots. Ustream changes the third party it chooses to license software from, presumably to one with fair-use caveats.
Interesting, still, to learn about companies like Vostream. That's got to be complicated work, and it's not even required by current law (as far as I know, I'm not a lawyer). I bet the business model is as much about charging for registration of copyrighted material as it is about charging for copyright violation detection. Now that I think about it, I bet Ustream doesn't actually pay anything for Vostream, and Vostream collects payments from content providers based on how many internet delivery sites agree to use it. Kind of a clever business plan.
posted by nathan v at 1:46 AM on September 4, 2012
I wouldn't watch an awards show if you paid me, but this has me so livid I think i can be seen from Mars.
This is why you can't trust robots.
posted by Mezentian at 4:40 AM on September 4, 2012 [1 favorite]
This is why you can't trust robots.
posted by Mezentian at 4:40 AM on September 4, 2012 [1 favorite]
The no-nothing child-s-school-report approach to blog-journalism.
Have you watched broadcast news lately? Blogs, love or hate, aren't the vanguard of journalism's decline.
posted by stavrosthewonderchicken at 5:22 AM on September 4, 2012
Have you watched broadcast news lately? Blogs, love or hate, aren't the vanguard of journalism's decline.
posted by stavrosthewonderchicken at 5:22 AM on September 4, 2012
Users of our paid, ad-free Pro Broadcasting service are automatically white listed to avoid situations like this and receive hands-on client support.
Fascinating. So "everyone's" pissed off because a free service failed? A convention decided to broadcast their signature event and couldn't bother to actually pay to make sure the broadcast worked?
And now they are pissed off that they got only half of what they didn't pay for?
Fury of the privileged indeed.
posted by eriko at 6:04 AM on September 4, 2012
Fascinating. So "everyone's" pissed off because a free service failed? A convention decided to broadcast their signature event and couldn't bother to actually pay to make sure the broadcast worked?
And now they are pissed off that they got only half of what they didn't pay for?
Fury of the privileged indeed.
posted by eriko at 6:04 AM on September 4, 2012
It's not hard to understand why the Hugo awards would go with one of the free streaming services - knowing in advance they'd cleared all copyrighted material with the rightsholders. This isn't some huge profit-making venture, if WorldCon turns a profit they use the money to partially reimburse panelists, staff and volunteers, who all pay to attend.
posted by mediareport at 6:46 AM on September 4, 2012
posted by mediareport at 6:46 AM on September 4, 2012
Point being, if you're going to offer a streaming service, free or paid, there's no necessity for you to go beyond what the law requires and do instantaneous takedowns without getting a complaint from a rightsholder first. If that's incorrect, someone please correct me.
posted by mediareport at 6:48 AM on September 4, 2012
posted by mediareport at 6:48 AM on September 4, 2012
Note the (hastily written and likely to be refined?) updated language on the Ustream apology page:
Users of our paid, ad-free Pro Broadcasting service “NOTE: UPDATED CLARIFICATION” ‘and those free broadcasters who notify Ustream in advance they have copyrights permissions (Ustream’s messaging to our broadcaster community how this process works is inadequate. We are resolving this now)’ are automatically white listed to avoid situations like this and receive hands-on client support.posted by nobody at 9:38 AM on September 4, 2012
mediareport, I believe there are two motivations for using content matching:
1. It helped Google fend off Viacom's billion-dollar lawsuit:
1. It helped Google fend off Viacom's billion-dollar lawsuit:
To a large extent, the case addressed past conduct, as Viacom said it was not seeking damages for any actions since Google put in its filtering system, known as content ID, in early 2008.However...
But Michael S. Kwun, a lawyer at Keker & Van Nest who previously worked at Google, said the decision would ensure that Internet companies were not legally required to develop such a system and could expect legal protection as long as they took down content when copyright holders complained. “I have no idea how much money YouTube spent on developing its content ID system, but if that was required for any new start-up, you wouldn’t see any,” Mr. Kwun said2. It can generate revenue. Content owners can be offered the choice of blocking matching videos, or monetizing them.
However, over the last two years, Content ID has been naturally migrating from an anti-piracy tool to a marketing and monetization tool. From the beginning, rights holders had the ability to leave unauthorized uploads up and monetize them with ads. The majority of partners now do so, said King. Those who leave user videos up have seen their overall views more than double.posted by jjwiseman at 4:36 PM on September 4, 2012 [2 favorites]
However, over the last two years, Content ID has been naturally migrating from an anti-piracy tool to a marketing and monetization tool. From the beginning, rights holders had the ability to leave unauthorized uploads up and monetize them with ads.
When one of my vids first got monetized that way, I took it down. Where the fuck was my cut?
posted by ChurchHatesTucker at 5:25 PM on September 4, 2012
When one of my vids first got monetized that way, I took it down. Where the fuck was my cut?
posted by ChurchHatesTucker at 5:25 PM on September 4, 2012
Now content ID is hitting the Democratic National Convention. Right now, I can't view their YouTube video because it "contains content from WMG, SME, Associated Press (AP), UMG, Dow Jones, New York Times Digital, The Harry Fox Agency, Inc. (HFA), Warner Chappell, UMPG Publishing, and EMI Music Publishing, one or more of whom have blocked it in your country on copyright grounds."
I wonder if Congress will now become more aware of the problem.
posted by grouse at 8:57 PM on September 4, 2012
I wonder if Congress will now become more aware of the problem.
posted by grouse at 8:57 PM on September 4, 2012
I wonder if Congress will now become more aware of the problem.
Are you a betting man?
posted by ChurchHatesTucker at 9:23 PM on September 4, 2012 [1 favorite]
Are you a betting man?
posted by ChurchHatesTucker at 9:23 PM on September 4, 2012 [1 favorite]
Wired on last night's takedown of the DNC stream on YouTube.
posted by ChurchHatesTucker at 6:51 AM on September 5, 2012 [1 favorite]
posted by ChurchHatesTucker at 6:51 AM on September 5, 2012 [1 favorite]
I wonder if Congress will now become more aware of the problem.
Are you a betting man?
Both Major Parties Are In 'Vigorous' Denial About The Need For Copyright & Patent Reform
posted by homunculus at 12:11 PM on September 5, 2012 [3 favorites]
Are you a betting man?
Both Major Parties Are In 'Vigorous' Denial About The Need For Copyright & Patent Reform
posted by homunculus at 12:11 PM on September 5, 2012 [3 favorites]
The Algorithmic Copyright Cops: Streaming Video’s Robotic Overlords
posted by homunculus at 9:25 PM on September 6, 2012 [1 favorite]
posted by homunculus at 9:25 PM on September 6, 2012 [1 favorite]
UStream will broadcast the Hugo Awards in full this Sunday
posted by Artw at 9:34 AM on September 7, 2012 [1 favorite]
posted by Artw at 9:34 AM on September 7, 2012 [1 favorite]
UStream will broadcast the Hugo Awards in full this Sunday
Have they done this in previous years? I could see this ending up with higher "ratings" than otherwise.
posted by ChurchHatesTucker at 1:24 PM on September 7, 2012
Have they done this in previous years? I could see this ending up with higher "ratings" than otherwise.
posted by ChurchHatesTucker at 1:24 PM on September 7, 2012
(And jscalzi is also livetweeting it, which is entertaining so far.)
posted by restless_nomad at 5:08 PM on September 9, 2012 [2 favorites]
posted by restless_nomad at 5:08 PM on September 9, 2012 [2 favorites]
« Older J'akuze | Crossing streams in the '80s Newer »
This thread has been archived and is closed to new comments
posted by Renoroc at 3:26 PM on September 3, 2012