Original URL: http://www.theregister.co.uk/2006/04/06/world_of_warcraft_case/

This means Warcraft!

It's time to call off the lawyerbots

By Mark Rasch

Posted in Media, 6th April 2006 11:23 GMT

Comment A recent World of Warcraft case involved a WoW book by Brian Knopp that was being sold on eBay. It resulted in automated takedown notices by "lawyerbots" and shows how the legal process today can end up silencing legitimate uses of trademarks and copyrights.

One staple idea of 1950s science fiction movies was of robots that take over the world. For example, in The Day the Earth Stood Still, a robot named Gort was poised to destroy the Earth, awaiting commands from Michael Rennie's alien, Klaatu. War of the Worlds (any version) similarly saw automated robots poised to take over the planet. But by the beginning of the 21st Century, a new threat emerged - attorneys. Now take these two and merge them, and you have a new scourge - the automated attorney. The lawyerbot.

Increasingly, legal notices, threats of litigation, and other legal process are being issued - and acted on - not by bespectacled gentlemen and women in crisp tailored suits, but by autobots, robots trained for litigation. These lawyerbots threaten to, like their metallic counterparts, take over the world. They must be stopped. Come to think of it, perhaps their human (semi-human) counterparts need to be stopped as well.

World of Warcraft lawyerbot

Brian Kopp lives in Bronson, Florida where he became something of a gamer on his purchased copy of the online multiplayer video game World of Warcraft. In fact, he became so good at the game, he achieved level sixty as a night elf rogue (this apparently means something to gamers). He decided to share his wisdom with others by writing a book containing hacks, cracks and cheats as well as tips, techniques and experiences for players playing the game. While this was an "unauthorised" book (with no permission from the Warcraft publisher) there was no evidence that Kopp's book contained any of the publisher's intellectual property. There were no bits of code in the book, references to source code, and only a few screen shots of the game taken from a third party website. In fact, the book clearly stated that it was "unauthorised" and that there was no affiliation between the book and the copyright holder. Fair enough.

Then Brian began to sell - or more accurately to try to sell - his book on eBay. That's where the trouble began. Shortly after the book became available for sale, eBay began to receive probably automated "takedown" notices telling them that Kopp's book violated the intellectual property rights of the various copyright holders. One after another, the "lawyerbots" kept notifying eBay about the illegality of Kopp's actions - all under penalty or perjury. "I, ROBOT.. do solemnly swear…"

Now these "takedown" notices are due to a provision of the Digital Millennium Copyright Act. You see, ISPs and other services that host content for others have had a legal problem. Third parties would host, post or otherwise display content that might infringe the copyrights of others. The parties injured by the postings would sue not only the poster, but also the ISP that was hosting the information. Under then existing copyright law, the host was making a "copy" in the cache of the infringing work, was doing so for commercial purposes (they were either being paid to host content, or drive traffic, or advertisements or whatever), and were contributing to the infringement. Therefore, hosting companies could possibly be held liable for content under their control unless they did something.

Congress stepped into the fray by passing particular provisions of the DMCA. These essentially gave these hosting companies immunity from liability for copyright infringement if they followed certain procedures. A party that felt that its intellectual property rights were being infringed would have to send the hosting company a notice - under penalty of perjury - swearing that they owned a copyright to the work, that the material hosted infringed the copyright, that it was not authorised or licensed, and the that posting was causing some infringement or damage - so help me God. This takedown notice has to be in writing and signed (electronic writings and signatures are ok).

When a hosting company gets a takedown notice, they are supposed to contact (or at least attempt to contact) the poster. If the poster doesn't respond, the host must take the offending material offline, and in fact is given immunity for doing so. If the poster swears under penalty of perjury that the materials don't infringe, this then gets transmitted back to the putative copyright holder, who must respond in 14 days. If there's no response, the materials stay up. If there is a dispute about it, then we go to court. A court can issue an injunction to remove the works, or can decide to let them stay up.

The system is intended to represent a balance between the rights of copyright holders, the needs of ISPs, and the ability of people to make fair use of copyrighted materials. Several things however conspire to alter this balance in favor of copyright holders.

RoboLawyers

Many threats to companies, such as phishing attacks, spam, copyright and trademark infringement, occur with such frequency (particularly on well-known trademarks) that it is simply impractical to personally review each and every message, write a formal letter to every mail host and ISP, and then litigate the potential copyright infringement. So, many companies have automated the process of detecting and responding to potentially infringing materials. If you are the Great Amalgamated Savings and Loan Company, you might employ an automated tool to search for references to you on websites, auction sites, message boards, chat rooms, etc. The tool can then be programmed to identify (or attempt to identify) improper uses of your name, trademark, copyright, trade secret, or other intellectual property rights. All well and good. In fact, if you have valuable intellectual property, you have a duty to protect it, and to be knowledgeable about potential infringement.

These programs can then go one step further. You can automate the process of sending out letters to the web host to take down the offending works. Now there is no indication that that is what happened in Mr Kopp's case. However, his eBay auction generated a slew of takedown notices from various parties. As soon as he reposted the auction, it generated a new takedown notice. Human lawyers are generally not that efficient. So these autonomous agents may in fact be the ones generating these takedown notices.

Chilling effect

One of the problems with these automated takedown notices is the fact that most ISPs will send a perfunctory notice to the last email address of the poster (if they even have that) and then just remove the putatively offending material. In Kopp's case (PDF), under eBay's Verified Rights Owner or "VeRO" program, eBay went even further - not only removing the allegedly infringing materials, but also suspending Kopp's account. So Kopp could not sell ANYTHING - not just the Warcraft book. When he opened a new account, the takedown notices would come again, and the new account would be suspended. Most people - even those who don't infringe, or have a colorable claim of non-infringement, simply walk away, tail between their legs. Thus, by wallpapering the net with takedown notices, a copyright holder (or trademark holder, or person claiming any kind of damage, breach, infringement, or improper use) can effectively remove all kinds of content from the web. And there are few if any consequences to guessing wrong. At worst, the alleged infringer can send a letter back and get the content put back up. Nothing stops you then from either contesting the use in court, or just letting cyberlawyer send you another takedown notice! You won't hurt its feelings.

Problems with copyright and trademarks

One of the biggest problems with lawyerbots is their inability to think and discern - particularly in the area of infringement. One might make the same argument about human lawyers as well. You see, copyright or trademark infringement isn't really binary. A work doesn't either infringe or not – there are infinite shades of grey. A clear case of infringement might be where I copy the entirety of your copyrighted work and sell it as you and keep the money. Pretty black and white. But in most cases, even when I copy parts of your copyrighted work, it may not be an infringement. Courts will look at whether my actions deprive you of substantial revenue. Whether I am doing it for commercial or other purposes. Whether I have copied all or a substantial portion of your copyrighted work, or only a small fraction. Whether my copy is for educational, literary, or commentary purposes. Or even whether you actually have a copyright in the work at all.

Similarly, in the area of trademark, it depends on whether you have a legitimate mark, and how far it extents. A court will also look at whether my use of your mark creates a "substantial likelihood of confusion". And whether my use of your mark in some way diminishes or disparages your famous mark. Courts take testimony, hear arguments, study law and precedent, and eventually make a ruling. As far as I know, lawyerbots don't. So the lawyerbots' emphatic sworn statement that it has a good faith basis to believe (or, more accurately that the copyright holder has such a belief based upon the lawyerbot's representation) that the work is an infringement is based principally on the fact that there is something in the posting that offends the copyright holder. In most cases, this is enough to get the stuff removed, censored and censured.

In Kopp's case, Brian finally did reach a real live human - well, a lawyer. Kopp explained that his work was clearly noted as unauthorised, used none of the copyright holder's copyrighted works, and was intended for commentary and educational purposes (and sold commercially). The lawyer insisted that the works infringed (even though you could only really use the book in conjunction with a purchased copy of Warcraft) because Brian was using their intellectual property (copyright, trademark?) for commercial purposes, and "attempting to trade off the substantial good will" in the World of Warcraft brand.

A quick look at Amazon.com's online bookstore identified 2,689 listings for a search for the term "Microsoft Office". In fact, in the previous sentence, I just referenced Amazon - a trademark holder - in an article published here for commercial purposes. Under this rationale, all of the "Idiot's Guides" and "Dummies" books, as well as technical journals, articles and references are infringing. Note that the lawyer didn't claim that Kopp's book disparaged the copyright holder, that it made them look bad, or injured the trademark. He never claimed that the book created any confusion about its source. No claim was made that there were copyrighted materials in the book. No claim that any sales were affected. In fact, he never even said which IP right was infringed. Just that the book was essentially "about" Warcraft, and published without permission.

The letter ominously concludes: "You are not allowed to sell an unauthorised 'guide' that attempts to trade off the substantial good will and recognition that (trademark holder) has built up in connection with its World of Warcraft product. In addition, the EULA (End User License Agreement) prohibits using the World of Warcraft software for 'commercial purposes'. Your disclaimer that these guides are for 'educational purposes only' is ineffective. Please consider this a warning. If you continue with the aforementioned activities, we will have no (sic) other alternative but to review all legal remedies available to us including taking formal legal action to protect our rights."

So the infringement seems to have been writing a book and selling it. "Danger! Will Robinson!" Moreover, by relying on the commercial use provisions of the EULA, effectively the lawyer is arguing that a software review published for commercial purposes would be a "use" of the software, and therefore both a violation of the EULA and then a copyright violation. The lawyer went on to lecture Brian about Brian's ignorance of trademark and/or copyright law, and suggest that Brian search the internet for terms like "intellectual property," "trademark," and "copyright".

Of course, the company had no need to resort to "formal legal action". No need to go to court. Just keep those automatic DMCA notices coming in, and eventually eBay and others will suspend the accounts. But Kopp did one better and sued the copyright holders and those who sent the DMCA notices. You see, it is rare that the issuer of a takedown notice is called to prove that what he swore under penalty of perjury is, in fact, true. So there are few consequences to calling a non-infringing use infringing, and great benefits if you manage to get the materials you don't like removed.

When you automate the process you end up silencing legitimate uses of trademarks and copyrights.

As Klaatu explained about the robot Gort: "In matters of aggression, we have given them absolute power over us. This power cannot be revoked. At the first signs of violence, they act automatically against the aggressor. The penalty for provoking their action is too terrible to risk."

It's time to call off the robots. "Gort! Klaatu barada nikto!"

Copyright © 2006, SecurityFocus

This article originally appeared in Security Focus.

Mark D Rasch, J.D., is a former head of the Justice Department's computer crime unit, and now serves as senior vice president and chief security counsel at Solutionary Inc.