Investor’s Business Daily on Filtering

by on September 11, 2007 · 9 comments

Brian Deagon’s August 6, 2007 article in Investor’s Business Daily, August 6, 2007, “Technology Doomed To Failure, Some Critics Say,” includes some remarks about filtering worth thinking about. The assurance of the quoted critics is convincing, but they seem to be missing a good part of the picture.


One argument is an analogy to Napster:

With the collapse of Napster, which enabled the exchange of music files through a centralized computer system, programmers developed peer-to-peer file-sharing systems such as Kazaa. With P2 P, the content resides on the computers of millions of individuals who download the software that lets them share files across the Web. . . Developers will surely try to subvert any filtering system.

And the article goes, quoting Public Knowledge in describing one method of subversion:

And with the content inspection approach, pirates could subvert the system by encrypting their files so that filters can’t identify them. . . Most P2 P technologies today don’t encrypt their signal, but all of them could.

The success of this method in foiling filters, though, depends on whether any features of what the filter is looking for survive encryption. Most watermarks are being designed to do just this (they need to be able to survive analogy conversion, as well).

More importantly, though, is the fact that the legal context had changed since Napster. Developers can develop subversion tactics until they are blue in the face, but so long as their use is held to a gray or black market, well, this is something content can deal with. Because of Grokster, though, a “subvert filter” menu isn’t going to show up on YouTube backed by the might of Google.

A second set of concerns expressed in the article is that the filters could tag and block legal content. A filtering system set up to look for content registered by the copyright owner isn’t going to be collecting a whole lot of mom’s photos of junior, it seems to me. The tricky part will be clips and bits and quotes. This is a real problem, if eBay’s trademark enforcement system is a model to go by. But it seems like one of those bridges that an appropriate procedure for contested blockings could cross when it comes to it, not a show-stopper of a problem.

Then there are privacy concerns, which rest on a misapprehension of privacy rights.

And last but not least, the concern that effort on developing filters is being misplaced; instead, the content community should be adding features that let them compete with “free.” But “free” comes in different flavors, shapes, and sizes, and competing with “free” without the help of technology will be like trying to sell bottled water without the bottles.

  • http://www.techliberation.com/ Tim Lee

    Really? Watermarks are detectable even after a song has been encrypted? I would be very surprised if such a watermark existed. Do you have an example?

  • http://www.techliberation.com/ Tim Lee

    Really? Watermarks are detectable even after a song has been encrypted? I would be very surprised if such a watermark existed. Do you have an example?

  • http://metapundit.net/sections/blog metapundit

    >The success of this method in foiling filters, though, depends on whether any features of what the filter is looking for survive encryption.

    I haven’t read the article so I’m surmising that the filters we’re talking about are trying to determine the content of files in transit and P2P systems might fight back by encrypting the files.

    Given those sets of circumstances there is no way that watermarks in encrypted files can be detected. It doesn’t work that way – encryption algorithms are designed to prevent any information whatsoever about the cleartext from being ascertained by examining the ciphertext.

    Now it’s true that watermarks survive analog->digital->analog conversions. And they would certainly survive cleartext->encrypted->decrypted transformations since these are lossless. But if watermarks could be designed that could be detected in encrypted streams this would imply fundamental flaws in encryption algorithms that would open them up to all sorts of attacks… I think you’re misunderstanding the way the technology is used.

  • http://enigmafoundry.wordpress.com/ enigma_foundry

    Solveig:

    A watermark might survive encryption, that is be present in the content after decryption.

    However, the use of encryption will disguise that watermark in the encrypted content. That is to say: if content is encrypted, the watermark will be undetectable in the encrypted stream.

    ef

  • http://metapundit.net/sections/blog metapundit

    >The success of this method in foiling filters, though, depends on whether any features of what the filter is looking for survive encryption.

    I haven’t read the article so I’m surmising that the filters we’re talking about are trying to determine the content of files in transit and P2P systems might fight back by encrypting the files.

    Given those sets of circumstances there is no way that watermarks in encrypted files can be detected. It doesn’t work that way – encryption algorithms are designed to prevent any information whatsoever about the cleartext from being ascertained by examining the ciphertext.

    Now it’s true that watermarks survive analog->digital->analog conversions. And they would certainly survive cleartext->encrypted->decrypted transformations since these are lossless. But if watermarks could be designed that could be detected in encrypted streams this would imply fundamental flaws in encryption algorithms that would open them up to all sorts of attacks… I think you’re misunderstanding the way the technology is used.

  • http://enigmafoundry.wordpress.com eee_eff

    Solveig:

    A watermark might survive encryption, that is be present in the content after decryption.

    However, the use of encryption will disguise that watermark in the encrypted content. That is to say: if content is encrypted, the watermark will be undetectable in the encrypted stream.

    ef

  • catus1

    I appreciate the concern which is been rose. The things need to be sorted out because it’s not about the individual but it can be with everyone.
    real estate search

  • catus1

    I appreciate the concern which is been rose. The things need to be sorted out because it’s not about the individual but it can be with everyone.
    real estate search

  • catus1

    I appreciate the concern which is been rose. The things need to be sorted out because it’s not about the individual but it can be with everyone.
    real estate search

Previous post:

Next post: