YouTube forced to reveal user logs

[Gavin Scott]

Google must divulge YouTube log

Google must divulge the viewing habits of every user who has ever watched any video on YouTube, a US court has ruled.

The ruling comes as part of Google's legal battle with Viacom over allegations of copyright infringement.

Digital rights group the Electronic Frontier Foundation (EFF) called the ruling a "set-back to privacy rights".

The viewing log, which will be handed to Viacom, contains the log-in ID of users, the computer IP address (online identifier) and video clip details.

While the legal battle between the two firms is being contested in the US, it is thought the ruling will apply to YouTube users and their viewing habits everywhere.

Viacom, which owns MTV and Paramount Pictures, has alleged that YouTube is guilty of massive copyright infringement.

The UK's Premier League association is also seeking class action status with Viacom on the issue, alleging YouTube, which was bought by Google in 2006, has been used to watch football highlights.

Legal action

When it initiated legal action in March 2007 Viacom said it had identified about 160,000 unauthorised clips of its programmes on the website, which had been viewed more than 1.5 billion times.

Following the launch of its billion-dollar lawsuit, YouTube introduced filtering tools in an effort to prevent copyright materials from appearing on the site.

We urge Viacom to back off this overbroad request

The US court declined Viacom's request that Google be forced to hand over the source code of YouTube, saying it was a "trade secret" that should not be disclosed.

But it said privacy concerns expressed by Google about handing over the log were "speculative".

Google's senior litigation counsel Catherine Lacavera said in a statement: "We are disappointed the court granted Viacom's over-reaching demand for viewing history.

"We will ask Viacom to respect users' privacy and allow us to anonymise the logs before producing them under the court's order."

The ruling will see the viewing habits of millions of YouTube users given to Viacom, totalling more than 12 terabytes of data.

Viacom said it wanted the data to "compare the attractiveness of allegedly infringing video with that of non-infringing videos."

'Erroneous ruling'

Leading privacy expert Simon Davies told BBC News that the privacy of millions of YouTube users was threatened.

He said: "The chickens have come home to roost for Google.

"Their arrogance and refusal to listen to friendly advice has resulted in the privacy of tens of millions being placed under threat."

Mr Davies said privacy campaigners had warned Google for years that IP addresses were personally identifiable information.

Google pledged last year to anonymise IP addresses for search information but it has said nothing about YouTube data.

Mr Davies said: "Governments and organisations are realising that companies like Google have a warehouse full of data. And while that data is stored it is under threat of being used and putting privacy in danger."

The EFF said: "The Court's erroneous ruling is a set-back to privacy rights, and will allow Viacom to see what you are watching on YouTube.

"We urge Viacom to back off this overbroad request and Google to take all steps necessary to challenge this order and protect the rights of its users."

The body said the ruling was also potentially unlawful because the log data did contain personally identifiable data.

The court also ruled that Google disclose to Viacom the details of all videos that have been removed from the site for any reason.

Story from BBC NEWS:
http://news.bbc.co.uk/go/pr/fr/-/1/hi/t ... 488009.stm

Published: 2008/07/03 15:58:10 GMT

© BBC MMVIII

Not very happy to hear this, but one could argue YouTube had it coming.

[Jovis]

At first glance it appears bad, but I'm sure that somebody with more knowledge than me on this will assure me that it's not. I hope.

However, I agree, it's clear that they're on dodgy legal ground - but naturally, it's hard to control what gets uploaded - it's reactive moderation, after all.

[Sput]

It's going to be mighty hard for them to mine that data into anything useful in any short space of time. I wonder if Google is trying to overload the prosecution with data. I doubt they're going after people that watch the stuff using that data, though, as there's no search or HTTP referral data that would help them identify where visitors to certain clips came from (and it they were looking for it explicitly).

[marksi]

I suppose it could be supplied as a hard copy. Not very environmentally friendly but a damn sight harder to dig through...

[nidave]

I suppose it could be supplied as a hard copy. Not very environmentally friendly but a damn sight harder to dig through...

Thats would be very funny to watch - poor postman.

[Mr Q]

It's going to be mighty hard for them to mine that data into anything useful in any short space of time. I wonder if Google is trying to overload the prosecution with data. I doubt they're going after people that watch the stuff using that data, though, as there's no search or HTTP referral data that would help them identify where visitors to certain clips came from (and it they were looking for it explicitly).

Apparently Viacom wants the data to prove that 'illegal' content (stuff that they, among others, hold copyright over) is more popular than home-made or authorised stuff. If they know what people are watching, they can demonstrate how seriously Google is taking the issue of removing copyright-protected content from the YouTube service.

[Dr Lobster*]

a 12 terabyte log file is pretty huge. at work we run isa, squid and apache (which function as a firewall, web proxy and web server), i've looked at the log files from each of these and they all have (fundamentally) the same information in slightly different arrangements.

the length of a single request entry ranges from between 244 to 380 characters, this includes such information as user agent string, source and destination ip address, date/time, http response code and the url. i would say the logs from google are going to have similar information in them.

now, 12 terabytes is 13,194,139,533,312 bytes. if we say for the sake of argument that an average line in the log file is 320 bytes (and that 1 byte=character, it's unlikely to be unicode), including linebreaks then we have just over 41 billion lines of log file to analyse (41,231,686,041 to be precise).

i have to say, i think pulling all that together to *prove* their case that viacom content is more popular than anything else on the site is going to mean that they are going to have to identify every single request.