Slashdot Mirror
Schmidt Says YouTube 'Very Close' to Filtering System
cnetfeed writes "Google CEO says an automated system will soon be available to track pirated content and prevent it from being uploaded to video sharing site. The system was supposed to be rolled out as early as last October, and the long delay in brining the technology online has resulted in ill will from companies like NBC and Viacom. 'Network executives accused Google of stalling so YouTube could reap the big traffic that professionally-created shows generate. Viacom filed a $1 billion lawsuit against Google last month and accused Google of massive intentional copyright infringement. "Ah Viacom," [CEO Eric Schmidt] Schmidt said. "You're either doing business with them or being sued by them...we chose the former, but ended up the latter." Schmidt took the opportunity to poke fun at Microsoft's assertion that Google's pending acquisition of DoubleClick may be a threat to fair competition. Other companies, including Yahoo and AT&T have also asked regulators to review the transaction closely.'"
This is great! Google has automated a way to look at videos and determine if they match a video that's already in existence and under copyright! That means they've solved a hard AI problem. I hope some day they open source their solution.
Apology to Ubuntu forum.
Maybe Google can turn Doubleclick into a better company--who knows? I am not drawing any conclusions yet, I want to see what they do with it.
and $1.6b down the "tube" when the visitors abandon the site because they cannot get their favourite show
there is no brand loyalty on the Internet when it comes to video sharing sites, he who has the content wins
and with all the popular shows gone because of this ID system why bother using YouTube when there are hundreds of video sharing sites (many not based in the US) that dont employ these tactics ? want freedom ? simply dont use US based services
Doubleclick ads were going to be there, regardless of who owned it. Google wanted the user data that Doubleclick had collected over the years, and they didn't want Microsoft to be able to buy it. Therefore they overpaid by billions of dollars.
The outcome was binary - either Microsoft was going to own Doubleclick, or Google was. Given the choice, I would much rather see Google do it. After all - the worst case scenario is that Doubleclick ads persist as they are... with a decent chance that Google is going to do something productive with Doubleclick's business model.
I may be completley wrong, but here's how I understand this so far..
If we can discount anything that looks at the actual audio or video content of the upload, and I believe we can since there just isn't an AI out there which can conclusively pick (for example) Captain Picard out of a video clip, then this must somehow use the textual metadata hooked to it. It'll have to sift through the tags and descriptions put there by the uploader, or possibly contextual data from sites that later embed the clip.
That would open some new cans of worms, though. First, it'd be easy to defeat, as we learned back when the old Napster suddenly didn't have anything by "Metallica," but there were tons of new songs from "Metallika," "Mettalicca," and "Metalligreed." Second, what if I record, say, a "C.S.I." parody? By rights I should be allowed to post it as such, will my file get flagged as lawsuit-bait and zapped because I used a copyrighted term in the description? What if I post an original film about firefighters that happens to use the word "heroes" in the title, which has nothing to do with the copyrighted TV series "Heroes?"
Slashdot Burying Stories About Slashdot Media Owned
For most of the videos, the following could easily be done...
p rint
Copyright holder notices a video that they hold the copyright to. They tell Google. Google checks the claims made, etc. etc. and presumably finds it valid. They make a hash of the video. They check their site for any other videos that match that hash, and remove those as well. They check any future upload and see if it matches the hash - if it does, it doesn't post it.
That leaves people getting around that by re-encoding, etc.
So in comes fingerprinting:
http://www.google.com/search?hl=en&q=video+finger
It's not exactly rocket science. Re-encode it? Zap a few frames? Fingerprinting tech laughs in the face of that sort of thing. The only two effective means of fooling fingerprinting tech are:
1. mangle the video so it can't possibly be recognized. Unfortunately, this means nobody can watch it without a special de-mangling player.
2. find out how the fingerprinting tech works, and make sure that only in those spots where it checks the original video, there's a difference. Of course any fingerprinting tech worth it's $$,$$$ will allow a seed value to change things around, and google rotates this once a week or however often needed once they realize people are getting around things in that way - and punish those users appropriately.
Not saying I agree with them doing it - though I find it hilarious in a sad way that if a copyright holder says "X is mine, please take it offline", that Google will do so - but not on Y which is the exact same video, or Z which is the exact same video getting uploaded a day later - but those are the ways they -can- do it.
I don't think anybody is suggesting that Google build an AI system that magically determines whether the copyright of a video lays with a third party and by means of technological ESP determines that said copyright holder did not consent to the upload - and I certainly don't think that Google is claming that they are either.