Slashdot Mirror
The Algorithmic Copyright Cops: Streaming Video's Robotic Overlords
thomst writes "Geeta Dayal of Wired's Threat Level blog posts an interesting report about bot-mediated automatic takedowns of streaming video. He mentions the interruption of Michelle Obama's speech at the DNC, and the blocking of NASA's coverage of Mars rover Curiosity's landing by a Scripps News Service bot, but the story really drills down on the abrupt disappearance of the Hugo Award's live stream of Neil Gaiman's acceptance speech for his Doctor Who script. (Apparently the trigger was a brief clip from the Doctor Who episode itself, despite the fact that it was clearly a case of fair use.) Dayal points the finger at Vobile, whose content-blocking technology was used by Ustream, which hosted the derailed coverage of the Hugos."
Geeta Dayal is a she. Just sayin'.
We have repeated cases of people going to court to dispute 'fair use', which shows that it is not well defined enough for humans to get right, let alone automated bots.
Lay down specific rules for 'fair use' and then you can write an algorithm to respect those rules.
(Just don't let RIAA/MPAA dictate the rules.)
The solution is to implement penalties for false takedown requests. Say, $25 per user per stream.
I'm sure he'll get right on it. All he has to do is get repeal past the Republican majority in the House and the permanent Republican fillibuster in the Senate.
Make no mistake, takedowns are always malicious, by their very nature. And the law that permits/demands it is even more so. I still hold on to the hope that someday our communications systems (internet, telephone, broadcast, etc) become robust enough to make all censorship impossible. Must destroy central control. That is our obligation.
“He’s not deformed, he’s just drunk!”
Unfortunately, this has only agitated people who already were against automated copyright filtering and DRM. It's like telling eskimos snow is cold. No, we'll have to wait until the MTV music awards are knocked offline by copyright bots before anybody who didn't already know about them gets wind of it.
Admit it. You post strawman arguments as AC so you get modded Insightful for refuting them, rather than Troll
A lot of technology is like OxyContin: it is very easy to abuse. The manufacturers/deployers make money and never suffer the negative effects. It's disingenuous to say that the technology is neutral and does not embody an business/political agenda. In this case allowing fair use would make the system much more complex, and might render it useless. For example if there were meaningful fines for false positives then those using this technology would have to act differently. Hell will freeze over before that happens.
Why is Snark Required?
You're not paying the copyright overlords. You're paying the video distribution system. If you're using the system for free, you can't expect them to take the lawsuit risk for you, so you shouldn't complain if they impose a stupid filter robot. So pay a fee (which probably gets rid of those ads people love to complain about) and show them you have permission to use the material you're broadcasting, so they can safely turn off the filter.
This is nothing new. "Clearance" has always been a major part of making movies and TV shows. (You know why the little kid in E.T. ate Reese's Pieces? They couldn't get permission to use Skittles.) Creative people have to work with the system, in part because it's the same system that allows them to profit from their work.
Mind you, I'm not defending the copyright overlords, with the legal sledgehammers and retroactive copyright extensions. But as fucked up as the system is, it's the one we've got, and the problems of dealing with it are nothing new.
What makes you think that Obama wants to repeal the DMCA? With the amount of support and money he is getting from Hollywood it is not surprising he is not mentioning copyright at all.
Negative moral value of force outweighs the positive value of good intentions.
The DMCA will probably never be overturned in the US, there is too much industry money behind it, and we know what feeds the political machine in the US.
If some third-party copyright trollbot interferes with the legitimate viewing of a webcast event, there has to be a law firm somewhere that, for the notoriety alone, would be willing to file a class action suit alleging damages of inconvenience and anguish on the behalf of thousands of viewers. Moreover, the broadcaster could sue for the costs of their broadcast that was interfered with. It costs real money to do a good quality webcast, trolls should be on the hook for diluting the value of a broadcaster's investment.
This technology was designed to find infringement.
I seriously doubt that. I think it was designed to find matching content and CLAIM it to be infringement while really having no means whatsoever to determine that.
now we need to go OSS in diesel cars
Takedowns have a legitimate purpose.
No, they seem to be abused time and time again. Go through the courts if you have an issue. You can't take shortcuts.
As an extreme example, what if it's child porn?
I like how people mention this as if it's self-evident. Honestly, not everyone has an irrational, insatiable desire to go after people who look at images or websites hosting them; some people would rather the actual perpetrators get caught.
But what does this have to do with takedown notices?
What if it's a bootleg of your favorite movie that just came out on DVD?
Why did you even mention this? Why would I care? Incidentally, ask a judge to order the content removed.
Whether something is infringing or NOT infringing, cannot be determined by matching content alone. Anything that only looks for matching content is intrinsically and fundamentally broken. Anyone who would design it that way and claim it to be correct is a liar and a fraud.
Correct. But that's exactly my point: the anyone who would etc is to blame and the reason for the blame is not because they designed the technology, but the way they use it (or claim it can be used) - is the second part of the logical conjunction that renders them liars.
Questions raise, answers kill. Raise questions to stay alive.
This is just a part of a much bigger problem, and that is who is responsible for the behaviour of rouge A.I.? Soon we'll have to face this problem head-on. With the military drones and robots being more and more autonomous, it's only a matter of time before DARPA comes up with the Terminator(TM), capable of autonomously deciding on the killshot. Then it's a matter of time before this machine makes a mistake killing a civilian or a war journalist and we'll find ourselves in the deepest legal shit humanity has faced since the Nuremberg Trials. You can't send a robot to prison, you can't charge it's maker, you can't lock up the author of the software. Yes, you CAN demand damages from the government/military but the crime still goes unpunished. So, should we grant the owners of such A.I.s a license to kill/threaten/arrest in autonomous nature and hope it doesn't turn ugly? And when it does, do we just say "shit happens, here's some money to make you feel better"?
I propose we make the people up-top personally responsible for such events. And there shall be peace on earth...
We live in an economic system based primarily on the concept of scarcity. Money is valuable because nobody ever seems to have quite enough of it, just like diamonds and gold. The law of supply and demand, there is more demand than supply.
The internet and the information age changes all that, its economics that is based on abundance. For the cost of a broadband connection we can give every human child access to the entire digitized archive of the collected works of the human race. The cost of doing this even 100 years ago would have been astronomical and beyond the reach of even the richest philanthropists. The old business models based upon everybody getting a percentage of the production costs break down then the production costs become zero. We have so thoroughly eliminated the nature of scarcity with information that our scarcity based economics, upon which we all depend for our basic survival, has deemed everything that abundance touches to be worthless. Rather than creating "utilitarian" value, it is seen as destroying "valuable" scarcity.
If you want economics based on abundance/gist-culture to actually work, then content effectively needs to be paid for in advance, with the risk of not-recieving, rather than our debt based culture of buy now pay later, where we only buy the hits after they have proven popular.
Copyright is about trying to maintain the concept of artificial scarcity in an environment where it doesn't naturally exist. But assuming that copyright has its place, there are checks and balances, but they don't all operate at the cost level and the speed we are used to in internet time.
The old school system of checks and balances are the courts and the legal system, You can spend a huge quantity of time, energy and money to argue your case and get a ruling from a professional judge. This works well when you have a small number of large, rich corporations with known fixed addresses and assets. Under the old school system, for things like copyright, most private individual use tended to simply fly under the radar and nobody made a big fuss about it regardless of the technicalities of the law. So disregarding morals, we are now down to game theory.
Now Mr Joe Blogs with his modern broadband connection could easily broadcast more copyrighted music than half a dozen fully funded commercial radio stations from a decade ago. Mr Joe Blogs has very little in sue-able assets beyond a bankruptcy order and by the time the cogs in the legal system have turned, he has managed to transfer more data than could fit on a supercomputer from 25 years ago. Joe blogs wasn't as issue when all he had was a mix tape and 5 friends, now he has a mix computer and 5 billion friends. Plus there are now a hundred million Joe Blogs on the internet doing exactly the same thing. So we have innocent until proven guilty, with a very high cost and high standard of proof required to enforce copyright, with the burden resting fully on the copyright holders.
So the copyright holders go to the politicians and say they cannot effectively enforce their business model, so they ask for the risk model of the checks and balances to be changed. Hence the DMCA, we want to be able to just send a letter and then shift the burden of proof onto Joe Blogs to argue his case, plus we want legal liability to rest with registered companies with static address who have something to lose if we have to take them to court. Guilty until proven innocent, but at least with the right of appeal.
But still the DMCA only works at the speed of the postal system, and requires human interaction, which is still orders of magnitude slower and more costly than "internet speed" which we all now mostly take for granted. Hence the advent of AI copyright bots that can at least operate at "internet speed", but the people who create them are on the payroll of the copyright holders and their definition of "failsafe" is to block content first and ask questions later. The checks and balances then start to operate, but they can only proceed a
If I remember correctly, these take down notices have a section where the issuer of the notice swears "under penalty of perjury" that the information on the notice is correct. When it turns out to be incorrect (or even when it isn't but no human ever checks the results from the bot), is that actionable? In a civil court? What is "penalty of perjury" exactly?
--
Stay tuned for some shock and awe coming right up after this messages!
hey now.. in today's radiant socialist future, ANYONE can identify with female if they choose.. it's just when they identify with male that they must be considered wannabe oppressors/terrorists/misogynists..
But why is there no attempt to borrow sentence construction and syntax from other languages when there is a clear benefit? So many languages have a gender neutral third person singular pronouns. For example Tamil has /avan/aval/avar/athu/ to mean /he/she/he or she/it/. Being Indian, I know Geeta is a typical Indian female name. But I cant tell a male first name from a female first name in many European, South American, Chinese and African cultures. And there are names used by both males and females in all languages. Gone with the wind had Ashley as a man's name. Agatha Christie wrote a whole mystery based on the idea Evelyn is a name used by both males and females. I think it was "Why didn't they ask Evans?" or Evil under the sun. Cant remember. There was an Indian MP by the name Kumari Anandan. Kumari with a short a is his home town used as first name. But with a long a, his first name translates as "Miss" in Hindi! He was assigned quarters along with female MPs and got routinely placed in railway sleeper coaches reserved for women!
English desperately needs a gender neutral third person singular pronoun. Time to coin a new word, something like "ce" to mean he or she. It could pronounced "see" half way between he and she.
Wish there is a bugzilla to file a ModReq on the English language.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact