AT&T Invests in Filtered Networking
Filtered Coward writes "Last summer, AT&T announced its intention to begin filtering copyrighted content at some point. The telecom has now bought a chunk of Vobile, whose core product is VideoDNA. "Like other systems of its kind, VideoDNA develops a unique signature from every frame of video. The signature is meant to be robust enough to survive various transformations and edits, and it can then be used to run matches against incoming content.' Vobile claims that VideoDNA is good enough to be used on video when transmitted over a network. 'Based on the complexity of the problem, we suspect that anything initially deployed by AT&T will fall far short of a robust P2P video filter. But should AT&T truly have its eyes on just such a prize, the company would be in a powerful position to impose its own policies on the entire US, since it owns major parts of the Internet backbone.'"
I thought telecoms were immune to certain types of litigation because they are neutral carriers of data. If a person makes a phone call or uses a bulletin board to commit a crime, the teleco is not part of the conspiracy. They are neutral. If AT&T starts filtering out "criminal" activity (and what jury of peers determined that anyhow?!), then are they giving up their neutral status? If they try to filter any material, will they be liable for all the material that inevitably slips through their net?
Also, how do they pick out copyright material for which a license has been granted compared to material that is "criminal" activity?
This is AT&T. They don't distinguish, they just give it all to the NSA, as demonstrated by the lawsuits filed by the EFF and the whisteleblower who revealed the taps on core fiber optic backbones.
The NSA, now, has fairly good tools. There's a fascinating tool from a company called Sandstorm that re-assembles network traffic into its distinct streams and does quite a good job of re-assembling email and web transactions. Given a remote opportunity to do a man-in-the-middle SSL key replacement, or simply steal the SSL or SSH keys from the serving host (with or without a subpoena), such tools could doubtless do quite a good job of intercepting transmissions seamlessly. And innocent folks aren't bothered to go to that level of protection, such as using obscure languages or real one-time pads.
Like the phone company's wilingness to tap phone conversations from the telephone offices, undetectably, because it's merely duplicating the digital bits and sending them to whomever they care to send them to, such monitoring constitutes a massive risk to the innocent for political and illegal monitoring. We see what such monitoring and related censorship does in China right now: we need to be extremely wary of it occurring here with such tools casually accepted.