The anatomy of an online copyright claim
This year’s World Science Fiction Convention (Worldcon) in Chicago was interrupted by a robot attack. On Sunday, Sept. 2, a live webcast of the Hugo Awards, one of the two top honours for science fiction, was cut off when a bot from the video streaming service Ustream detected alleged copyright infringement. The broadcast was terminated at 7:43 p.m. Pacific time in the middle of an acceptance speech by author Neil Gaiman. The allegedly offending material was a clip from the Doctor Who episode for which Gaiman won the award.
How did this happen? Well, Ustream’s founder, Brad Hunstable, posted an apology the next day, in which he admitted that the bot was too sensitive and the process for getting a copyrighted broadcast approved was not clear. The convention had permission from the copyright holders to air the clip, according to organizer and Hugo-winning editor Cheryl Morgan, and many have claimed that in any case the use of the clip should have been protected by the “fair use” doctrine in copyright law. The Hugo webcast was not protected against being on the bot’s whitelist (or anti-spam filtering software) because of the obscure procedures.
The “fair use” doctrine exists to prevent legitimate uses of copyrighted material from charges of infringement. The use of excerpts of copyrighted work for criticism or academic commentary, for instance, is protected, as is the use of material for parody. There’s no clear-cut way to know in advance how a court will rule on a fair use claim, but using a small extract from a TV episode for a presentation of an industry award falls well within the generally acceptable bounds.
The convention organizers were most likely justified in using the Doctor Who clip in their ceremony: it violated neither the laws of copyright nor the terms of service for Ustream. However, the copyright bot had no way of knowing this. It cut off the webcast of the Hugo awards and actually banned the convention’s Ustream account for showing something that they were entirely within their rights to show.
Because this all happened in real time, there was no way to address the error. The ceremony was on the Sunday night of a long weekend, and according to Morgan the convention’s Ustream account “did not rate contact with an actual human being.”
Even if the convention had managed to get in touch with one of Ustream’s administrators, the process of investigating the claim and unbanning takes time, and the ceremony cannot be put on hold while the webcast is sorted out. Even though Worldcon was vindicated in the end, their right of fair use has been effectively abrogated.
This is without any deliberate malice on the part of the Ustream people. They apologized and disabled the copyright bot until it can be more accurately calibrated. They don’t actually want to inconvenience their users, and the copyright bot’s purpose is to prevent, for example, people using the service to broadcast entire episodes of Doctor Who illegally. The problem is with the way their incentives are structured.
As an ad-supported service, Ustream cannot afford to get a reputation (just or unjust) for allowing copyright infringement on their site. Individual users are not worth much to them because end users are not their customers. We see this same dynamic at work on Facebook and other sites. So long as no single action ad-supported services like these provoke a mass exodus, enough users will stay that there is essentially no cost to implementing security measures that are troublesome and intrusive for the end user.
As a result, Ustream’s fair use-denying practices—along with Facebook’s privacy-denying practices—continue more or less unchecked. Even if the new copyright bot is more accurate, that doesn’t change the fact that it is designed to take down suspicious webcasts before any wrongdoing has even been established. Ultimately, this is the price of free services on the Internet.