Judge Orders Piracy Trial To Test IP Address Evidence 321
another random user sends word of a case in Pennsylvania District Court in which Judge Michael Baylson has ordered a trial to resolve the issue of whether an IP address can identify a particular person. The plaintiff, Malibu Media, has filed 349 lawsuits against groups of alleged infringers, arguing that getting subscriber information from an ISP based on an IP address that participated in file-sharing was suitable for identification purposes. A motion filed by the defendants in this case explains "how computer-based technology would allow non-subscribers to access a particular IP address," leading Judge Baylson to rule that a trial is "necessary to find the truth."
"The Bellwether trial will be the first time that actual evidence against alleged BitTorrent infringers is tested in court. This is relevant because the main piece of evidence the copyright holders have is an IP-address, which by itself doesn't identify a person but merely a connection. ... Considering what's at stake, it would be no surprise if parties such as the Electronic Frontier Foundation (EFF) are willing to join in. They are known to get involved in crucial copyright troll cases, siding with the defendants. We asked the group for a comment, but have yet to receive a response. On the other side, Malibu Media may get help from other copyright holders who are engaged in mass-BitTorrent lawsuits. A ruling against the copyright holder may severely obstruct the thus far lucrative settlement business model, meaning that millions of dollars are at stake for these companies. Without a doubt, the trial is expected to set an important precedent for the future of mass-BitTorrent lawsuits in the U.S. One to watch for sure."
Re:Responsibility? (Score:5, Informative)
No, you aren't, or at least not in the U.S. You are only an accessory if it can be shown that you had actual knowledge that the person who borrowed your car intended to rob a bank. If someone asks to borrow my car and then, without my knowledge, uses it to rob a bank, I am not an accessory.
Similarly, you are only an accessory if you knew that loaning him or her your car would help him or her in committing that crime or evading capture in some way. If somebody tells me he or she is thinking about robbing a bank and then, in a separate conversation, asks to borrow my car to go get milk, I am not an accessory even if he or she then robs the bank using my car.
You have to have not only knowledge of the crime, but also intent to aid in the commission of the crime.
Re:IPs parallel the discoverable world (Score:5, Informative)
MAC authentication is absolutely, literally, worthless from a security standpoint if you are using WPA2. Anyone who has the capability to crack WPA2 will necessarily have the ability to impersonate your MAC-- it is, I believe, a requirement to mount an attack against WPA2 in the first place. The fact that you have MAC auth turned on would probably not even be noticed by an attacker, and if it were, it would take all of about 5 seconds to get around.
Re:Responsibility? (Score:4, Informative)
Re:IPs parallel the discoverable world (Score:5, Informative)
In fact, if a person wanted to be really nasty about it, the following would be trivial to do:
1.) I passively monitor your WLAN in the evening.
2.) In the morning you leave for work, taking your laptop with you.
3.) I assign YOUR mac address to my pc and go about my illicit business.
Police come knocking on your door, check log files if your router has them, and right there in the logs is YOUR mac address from YOUR laptop correlated with the illegal activity.
Anyone who understands wireless networking, even a little, should know that the thought of an IP address being considered legal proof of identity is an absolutely TERRIFYING concept.
Re:IPs parallel the discoverable world (Score:4, Informative)
While the MAC address of a NIC can be tedious to alter
1.) Boot a backtrack iso
2.) Run macchanger -r eth0
There you go, you're now operating under a randomly generated MAC address.
Not too tedious, IMHO.
Re:IPs parallel the discoverable world (Score:4, Informative)
Re:IPs parallel the discoverable world (Score:5, Informative)
Run that through a database containing everyone on the planet, and you get a few million positive hits, all but one an error.
No you won't. There are 13 standard Loci with something like 10 Alleles or more at each marker. So that is something like the chance of a "random" match as one in 10^-12. This is both correct and wrong. First many of these 13 markers have more than 10 alleles and the provability is closer to something like 10^-15. Its wrong in that its not random, you share about 50% of these markers with your father for example. Even population wide this does reduce the randomness. Then there is a birthday paradox. But that does not apply in this case since you are matching the database to a given profile. So with 7 billion humans in the database, chances are that there is just one hit. Not millions. You would be very lucky to get more than one.
When comparing to a 100 suspects that are not related (remember the profile will tell us if they are related.) You are more like 99.99999999% sure. Even far more than that.
Yes this is directly related to my day job.
Re:IPs parallel the discoverable world (Score:5, Informative)
"When comparing to a 100 suspects that are not related (remember the profile will tell us if they are related.) You are more like 99.99999999% sure. Even far more than that. "
Wrong. The reliability of the person doing the testing accurately is not anywhere near 99.99999999% or even 99.999% and represents the absolute maximum assurance the test can provide. That is comparable to saying something weighs 1.34545g when your scale is only accurate to +/- .1g.
The lack of randomness does not make DNA profiling a better indicator, it skews the odds the other way. It proves that there are relationships in these markers. If I have a one byte binary number you can say that there are 2^8 possible numbers so the chances of a randomly picked number matching mine are 2^8. But the moment that number has a meaning the uniqueness of the indicator drops. If it is human readable English text then there are only 96 possibilities and my random selection now has a 1 in 96 possibility of matching. If it was a "random" keypress the odds become much better and a simple number can no longer express the odds because some numbers are more probable than others, for instance if my random key is a home row key the odds are dramatically better than 96 to 1. More like 20 to 1 and even within the home row some keys are more likely than others.
The point being, while we suspect these markers are very unique, there definitely have not been any studies on a sample set nearly large enough to assert a 1 in 99.99999999% probability with any degree of confidence. Those type of odds assume there is no relation between these markers and any relation can drop the real probability by several orders of magnitude.