News for hackers. Stuff that matters.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jedberg on Monday June 18, 2018 @12:24PM (#17339478)
I saw a great "joke" the other day. It said, "If Shazam or Siri can't identify a song for you, make a quick video and upload it to YouTube. Within a few seconds you'll get a copyright takedown telling you the name of the song".
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by reificator on Monday June 18, 2018 @12:45PM (#17339650)
I've seen nature videos where there's no sound but the birds chirping, and YouTube is still able to hear someone playing All-Star on Mars and flag the video. It's really incredible tech.

EDIT: The article actually mentions some similar examples: http://c4sif.org/2012/02/youtube-identifies-birdsong-as-copy...

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jedberg on Monday June 18, 2018 @01:01PM (#17339773)
I know the guy who created ContentID. He's a good guy and only had good intentions, namely to keep YouTube in business by identifying legit copyright issues. At the time, a human had to review all the claims.

This is a good example of how AI can go bad.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by tzs on Monday June 18, 2018 @03:30PM (#17340998)
Looking at the DMCA requirements, it doesn't appear like it would require a lot of human intervention.

1. Copyright owner provides written notification of the alleged infringement.

The notification has to provide reasonably sufficient information to allow Google to locate the alleged infringing material; provide contact information for the claimant; include a statement that the claimant has a good faith believe that the use is not authorized by the copyright owner, its agent, or the law; and a statement that the information the claim is accurate, and that the claimant is authorized to act for the owner.

2. Google then has to remove or disable the material, and take reasonable steps to notify the uploader of the issue.

If the uploader does not make a counter notification, it looks like that's the end of Google's obligations.

If the uploader files a counter notification, Google has to notify the claimant. If the claimant does not file a lawsuit against the uploader within 14 days, Google has to restore the material.

Note that nothing in here seems to require Google to actually look at the material and decide if they think it is infringing. Nor does it require them to proactively seek to find infringing material and notify content providers or block such material. They just have to check that the claimant (and the uploader if they get a counter notice) have included the things in their notice that the law requires. They should be able to automate that for most claims.

The ContentID system seems to be going far beyond what DMCA requires them to do to keep their safe harbor. I wonder why they are doing that? Maybe because they want to also be content sellers, and so want to keep the content providers happy with them?

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by mjevans on Monday June 18, 2018 @04:24PM (#17341407)
The problem is a lack of liability for the content creators. Confusion can happen (such as particular renditions of a work who's basis is public domain information (IE a classical music score)), but willful lies should be punished with actual consequences.

Right now things are stacked entirely against one class of content creators (video creators). Additionally if I'm a public institution or someone else that's against having ads I can't control what happens during the contest period* (or so I've heard, it's never happened to me).

As someone that has posted a few 'how to do X' in software videos before, who believes that ads should be far more limited (like the old soap operas / PBS sponsors at the end of programs), I'd rather my videos were in 'hold on publishing limbo' while not under my control.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by seandougall on Tuesday June 19, 2018 @11:16PM (#17352292)
In the case of classical music, there are two things that need to be clarified:

1) Classical music is still being composed today, and there’s a huge volume of post-1927 work that is very much under copyright in the US (excessive though that may be IMO).

2) Even if a composition is PD, the recording is absolutely copyrightable if it’s post-1927, and the arrangement may be as well. This isn’t a matter of confusion; it’s an actual piece of intellectual property.

Not to say that any of this is right; I believe fair use protections need to go much, much farther than they do.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by paulie_a on Monday June 18, 2018 @05:19PM (#17341788)
Will fill abuse of enforcement should invalidate the copyright entirety...it becomes public domain. Repeated violation revoke all copyrights under your control.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by yason on Monday June 18, 2018 @04:41PM (#17341530)
The ContentID system seems to be going far beyond what DMCA requires

AFAIK the content id system was built to monetise works on YouTube. The system must know where a song or a video is being included to direct advertising money to the original author if s/he wishes to participate in that scheme.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by sp332 on Tuesday June 19, 2018 @07:06AM (#17345488)
ContentID does go far beyond DMCA requirements. I think for a while, some big companies has carte blanche to delete any video. I guess narrowly winning a billion-dollar lawsuit against them will scare a company that way.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by reificator on Monday June 18, 2018 @01:06PM (#17339824)
I'm not demonizing the people making it, or even necessarily the people demanding it be made.

The DMCA was the beginning of the end of a free and open internet, and now it's accelerating. I just hope the next generation understands what they're missing under the spectre of copyright law.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by dangrossman on Monday June 18, 2018 @07:02PM (#17342331)
Prior to the DMCA, ISPs had full liability for copyright infringement committed by their customers online. Small ISPs were being sued out of existence, and it was only a matter of time before nobody would be willing to run an ISP at all. The DMCA (specifically Section 512) was the thing that saved the free and open internet by introducing limited liability for ISPs.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by reificator on Monday June 18, 2018 @08:59PM (#17342946)
That's fair and I don't want to downplay that aspect. But that's not the only thing the DMCA does.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jimbo999 on Tuesday June 19, 2018 @02:12AM (#17344192)
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by amyjess on Tuesday June 19, 2018 @11:59AM (#17347860)
While the DMCA has been abused by large corporations, it has been invaluable in helping protect the little guy.

It's been a huge boon to independent content creators that you can just send somebody a DMCA takedown notice and they'll comply. Before the DMCA, you had to pay out the nose for a lawyer, and you'd be in for a long and very, very expensive legal process. Your average photographer can't afford that, so copyright enforcement was solely a privilege of the rich prior to the DMCA.

Here's an example. I know someone who shared pictures of herself in a private women-only group. Someone who ran a really disgusting cyberbullying page on Facebook had a mole in the group and posted the pictures publicly while making fun of her appearance and calling her slurs. Since the photographer was a friend of hers, the photographer simply sent Facebook a bunch of DMCA notices telling the cyberbullying page to take down the pictures. Facebook complied in under an hour. The cyberbullying page tried to post the pictures again; more DMCA notices were sent, and the whole page was unpublished.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by simion314 on Monday June 18, 2018 @01:25PM (#17339980)
Humans are to blame for using horrible AI/software. At least if you know your AI is that broken then have some human supervision.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by narrator on Monday June 18, 2018 @03:38PM (#17341049)
The real problem is when the task is no longer doable by Humans. Then the AIs have taken over and we can't turn them off. Imagine what happens when the AIs run our cars and food production, security, etc. We will be unable to turn them off and they will be unaccountable except to the people who set their policy which will be mired no doubt in layers of deep obscurity.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by simion314 on Tuesday June 19, 2018 @02:08AM (#17344175)
Or the problem where the developers will not know why the AI did something, say an AI decides to do some bad thing, the judge will ask the developer why and the dev will say: -I don't know, we feed them a ton of input and it worked fine on my machine and without test data so we shipped it
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by quikoa on Monday June 18, 2018 @02:09PM (#17340308)
Google prefers to automate everything. Having human supervision would affect the bottomline so is out of the question.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by Asooka on Monday June 18, 2018 @02:23PM (#17340437)
Well, let's say it takes 30 minutes to fully check for copyright compliance 10 minutes of video. This means that the speed of one person checking videos is 1/3 unchecked-minutes per minute, or 1/3upm. 400 hours of content are uploaded every minute to youtube, which makes for 24000upm. You will need to employ 96000 people 24/7 to check all videos. Of course, they can't work 24/7, so a more realistic figure is 150000 people. This isn't menial work, it requires legal knowledge and familiarity with copyright law, so you're looking at hiring entry-level law clerks, which is about 16$/hr or about 33k$/yr. It will cost Google about 5bln$/yr to check all uploaded content manually. Which seems like a lot but is like 10-20% of Alphabet Inc.'s gross profit, so not actually impossible.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by earenndil on Monday June 18, 2018 @02:51PM (#17340684)
Why not simply have humans review the videos that are automatically flagged? That would be a significant improvement over the current state of affairs, and would command a significantly more reasonable price.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by simion314 on Monday June 18, 2018 @03:22PM (#17340945)
Exactly, have the code extract the exact section of video that it considers problematic on left side and on the right side put the copyrighted work, the human will lose a few seconds to have a look.

Also I do not understand why Google does not use a reputation system, if a channel has a good reputation then take this in consideration in the automatic detection.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by Dylan16807 on Monday June 18, 2018 @04:32PM (#17341479)
At the very least do it for major accounts (4k hours a year). It's pretty ridiculous that pulling in hundreds of dollars won't get you even a 30 second review.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by tomhallett on Monday June 18, 2018 @02:54PM (#17340706)
The question is whether you are more worried about false positives or false negatives.

If you are only worried about false positives, then the humans only need to review the content flagged by the AI. This has 2 benefits: 1) the "starting number" would be much much lower than 400 hours of content per minute. 2) If a 10 minute video is obviously pirated, you don't need to watch the entire video to decide that.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by greglindahl on Monday June 18, 2018 @02:41PM (#17340603)
Are the only 2 choices (1) no human in the loop and (2) only humans in the loop?
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by IncRnd on Monday June 18, 2018 @11:16PM (#17343541)
There are some strong corrections that need to be made to your math.

1. Only a subset of uploaded videos need to be checked by humans, those that were flagged by the automated system.

2. Only a subset of the runtime of each flagged video needs to be checked. In other words, if a human decides within 12 seconds of a video that it is infringing, there is no reason to check the remaining portion of the video.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by mygo on Monday June 18, 2018 @03:22PM (#17340954)
today in order to do a Google search I had to prove to Google that I wasn’t a robot. And even though I did it right the second time, they still made me do it a third time. I just remember sitting there going, Google, I actually am a human and your AI sucks so if I have to tap on storefronts and cars one more time to do a freaking Google search then I’m just using Bing.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jstanley on Monday June 18, 2018 @03:45PM (#17341112)
It's partly checking that you're a human.

It's partly using you for free labour to train their next generation of SkyNet.

We've reached a point with CAPTCHAs (originally conceived as a way for humans to exert their will over machines) where the machines are now setting tasks for the humans, and the humans are carrying out the tasks without putting up too much of a fight.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by eftychis on Monday June 18, 2018 @08:04PM (#17342681)
That. CAPTCHAs are used as free labour. The only reason you don't see them everywhere is because people would stop using Google.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by reificator on Monday June 18, 2018 @09:18PM (#17343044)
We've done it. We've cracked the code.

This thread is complaining about the fully automated process used by Google to check for copyright violations, and how infeasible it is to hire someone to check each video.

You're talking about how CAPTCHAs are now being used to get free labour.

> Prove you're human: Does the following video contain Beat It by Michael Jackson?

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by eftychis on Monday June 18, 2018 @09:44PM (#17343156)
I am pretty sure they have considered it. Probably there is a legal hiccup laying here -- you either have reasonable suspicion that the material you show is copyrighted, or you acknowledge your algorithms and AI suck and you require crowd-sourcing. Kind of a lose-lose legally, and they would possibly introduce liability.

Recall the goal of Google is to avoid liability via the statement "we are pretty good and have really low false negatives", and have as few false positives as not to make the platform unusable.

Google does not care about content creators or propagation of knowledge (instructive videos) and ideas. Simple statement here. Google engineers might (at least the friends I have), but not as a whole organization.

But yeah crowd-sourcing is a good way to do a second sifting through results. I know it is being used, by Google and others, already to classify video sources.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by briandear on Monday June 18, 2018 @11:53PM (#17343705)
Why do site owners use Google CAPTCHA? Even Cloudflare uses it. I resent being forced to help Google when not even on a Google site. I won’t put any of my stuff on Cloudflare specifically because they force people to help train Google’s image recognition.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by icebraining on Tuesday June 19, 2018 @03:37AM (#17344494)
Why do site owners use Google CAPTCHA?

Because it's free, reasonably hard to beat by bots and reasonably low-effort for the user if they let Google track them (enabling the single-checkbox captcha).

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by dvtv75 on Monday June 18, 2018 @09:54PM (#17343205)
A few days back, to allow me to download the software I had just bought, the site kept putting up those damned pictures for me to click on. Street signs were the first three, I apparently got all of those wrong (even though I got them right), shop fronts was the next lot (that's a little harder to work through, I can't tell Chinese shop fronts from Chinese warehouses by text alone) and I had two of those, then more street signs, and finally cars.

It took longer to confirm I was human than it did to download the software. My partner came through to see if I was alright, after I was shouting abuse at the computer for an extended period. She sat through one of them with me, confirmed that I had it right, then swore and walked away when it decided I was wrong.

If I'd known that it was going to screw me over that much, I would have recorded it and put the video on YouTube.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by sverhagen on Tuesday June 19, 2018 @02:06AM (#17344168)
Would that have infringed on Google's copyright?
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jimbo999 on Tuesday June 19, 2018 @02:11AM (#17344189)
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by Jyaif on Monday June 18, 2018 @01:38PM (#17340094)
ContentID is not AI by any definition of AI, it's signal processing.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by LordDragonfang on Monday June 18, 2018 @01:59PM (#17340243)
I believe the implication is that youtube has replaced the human reviewers with an ML-based system.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by kazinator on Monday June 18, 2018 @02:17PM (#17340387)
ContentID isn't learning because it doesn't train to make inferences; it looks for exact signature matches on the audio, normalizing for alterations like pitch changes and equalization.

All sorts of computer applications that are not considered AI have replaced humans. For instance, databases replace clerks having to shuffle papers to and from filing cabinets.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by gnulinux on Monday June 18, 2018 @10:14PM (#17343293)
I think saying signal processing is AI is like saying the entire CS is AI, which is bizarre. There are all sorts of algorithms and disciplines in CS that are not AI. I think the commenter you're replying to was complaining about this.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by thomastjeffery on Monday June 18, 2018 @01:58PM (#17340241)
It's automated decision-making.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by kazinator on Monday June 18, 2018 @02:21PM (#17340421)
Any algorithm that produces a Boolean result is "automated decision making". You pop in the inputs, let the automaton run, and there is your yes or no.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by gowld on Monday June 18, 2018 @04:21PM (#17341385)
Yes, that's right. The boundaries of AI are qualitative, not quantitative. AI can be simplistic or complex, dumb or smart.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jexah on Monday June 18, 2018 @06:49PM (#17342256)
AI doesn't have to learn to be classed as AI. Source: Every video game ever made. Just look at deep blue. Most would consider that AI, but AFAIK there is zero machine learning involved.

Edit: I think I hit the wrong reply button, meant to reply to parent, haha.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by glogla on Monday June 18, 2018 @02:11PM (#17340327)
I guess he won't go down in history like the leaded gasoline and freon guy. But I guess it's more like Nazi Germany where if it wasn't him, someone else would have.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by kchr on Tuesday June 19, 2018 @03:11AM (#17344397)
Dr. Godwin, I presume? tips hat
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by avian on Monday June 18, 2018 @02:52PM (#17340693)
I got a copyright strike once for a video with no apparent reason. After closer inspection, I could in fact hear a radio playing somewhere in the background if I really turned up the volume. I found it incredible that YouTube’s algorithm was able to pick it up at that signal-to-noise ratio.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by local_yokel on Monday June 18, 2018 @05:23PM (#17341809)
For people with a signal processing background, this is actually a very trivial thing. The basic technique has been in use for probably a hundred years - radar being a classical example - nothing at all to do with Google. Basically, you can get a signal-to-noise ratio improvement proportional to the duration of the signal you're attempting to detect (in this case, audio), thus allowing you to detect very weak signals in the presence of strong noise or other unwanted signals. Look up "pulse compression" or "correlation detection" if you're interested.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by gnulinux on Monday June 18, 2018 @10:16PM (#17343306)
This is very true. In UC Berkeley, the very first linear algebra course all EE/CS freshman take (EE 16A) has a lab that does EXACTLY this. You match part of a song with some very noisy sound. If freshmen are taught to do this, it's easily something Google can do.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by rosstex on Tuesday June 19, 2018 @10:44AM (#17347124)
Go Bears. I weaseled my way out of the requirement for 16A (when it was transitioning to replace EE 20/40), and I kinda regret it now.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by kchr on Tuesday June 19, 2018 @03:12AM (#17344405)
Still, it is quite incredible for us laypersons :-)
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by sverhagen on Tuesday June 19, 2018 @02:12AM (#17344193)
If it's background noise, is it still a violation? Not suggestion an answer, just asking. Like, how (not in all countries) it's okay to get passers-by in your video shot and broadcast them uncensored?
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by teaspoons on Tuesday June 19, 2018 @06:30AM (#17345324)
if it can identify the song why can't it process it out of the soundtrack ?
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by dingo_bat on Tuesday June 19, 2018 @08:31AM (#17345946)
They don't care. Individual videos don't matter to them.
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by userbinator on Monday June 18, 2018 @08:26PM (#17342784)
I actually tried that and got a name of a song, and it wasn't a takedown but an agreement to monetise (i.e. add ads, which I didn't have much of a problem), except that it was the completely wrong song. Naturally, that song was also available to listen to in a few other YouTube videos, and the only thing that they had in common with mine was a generic-sounding section of 120bpm bass drum. The melody, the duration, everything else was different.

I disputed, explicitly pointing to the other ones which I assume was identified correctly and how they had basically nothing in common with my video, even explaining that I'd be perfectly fine with it being monetised if it was actually the right one, and was... (not) surprisingly denied. I have a feeling that either no humans were involved at all, or if one were, it did not listen or have any common sense. In the end, I just deleted the video.

Now that I think about it, perhaps reuploading with the (wrongly) detected name, letting it monetise, and show up in search results might be a better idea --- at least that will make it more likely for others to stumble across it, listen, and possibly leave comments like "this isn't X, it's Y"...

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by dyukqu on Monday June 18, 2018 @10:43PM (#17343422)
Honest question: do you guys and gals think YouTube is really able (does it even try?) to identify the songs that Shazam or Siri can not identify - at all? I presume that if a song is unidentifiable for Shazam (or another service like it), then it's not popular and "important" enough on YouTube too so it doesn't get caught OR it's a folk song from a non-Western country.

It's highly probable that you can get the name of such a song much faster by asking about it in the comment section.

Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by jedberg on Tuesday June 19, 2018 @12:40AM (#17343891)
I've actually tried this with popular western songs. The difference seems to be that YouTube is better at identifying songs with a lot of background noise. Shazam/Siri/Google need to have cleaner audio to be successful (I'm not sure why the Google assistant doesn't use the same tech as YouTube, but it doesn't seem to).
Re:YouTube’s Piracy Filter Blocks MIT Courses, Blender Video... by golergka on Tuesday June 19, 2018 @11:30AM (#17347580)
I wonder, how many laws would I break if I implemented a Song ID service based on youtube APIs that way?