Stop, in the Name of 'Bots

Nowadays, it seems as if more and more law enforcement is being done by machines. Unfortunately, they don't seem to be up to the job. And the humans don't want to take responsibility, either.

One example is provided by the traffic cameras that are being employed by revenue-starved cities around the nation. They shouldn't make mistakes: They automatically photograph cars that are speeding, or that run red lights, and then they take a picture and a computer issues a citation. Unfortunately, they turn out to be inaccurate, and often rigged to produce revenue for the contractors and cities who operate them, at the expense of people who are innocent.

Then there are the software 'bots that the entertainment industry is using to look for copyright infringers. The software, which crawls the Web the way search-engine software does, looks for files that appear to match copyrighted material: songs, movies, etc. When the 'bot finds the material, a human being is supposed to check the results, then notify the Internet service provider that a customer is infringing upon protected material and that the material must be removed. This usually leads to the ISP terminating the Internet account of the individual posting the material.

Except that it doesn't work that way. Here's what the Internet Service Provider Association says in a brief just filed against the Recording Industry Association of America:

The lists have not even been culled to eliminate items that should never have been included in the first place. While most of the works identified in Exhibit 1 appear to be songs featuring George Harrison, the notice also demands removal of a file labeled, in part, "John Lennon, Yoko Ono And George Harrison Interview.mp3." The notice further objects to a file entitled "Portrait of mrs. harrison Williams 1943.jpg." It even claims infringement by distribution of a file whose appalling title includes phrases such as "Nude Preteens and Young Teens Naked ... Brian is 14 and Harrison is 8."

It is highly unlikely that a single company owns the copyright both to George Harrison's songs and to child porn materials. The most likely explanation is that the list was generated by a "bot" that had been programmed to search for all files with the name "harrison" in them -- and that the results of its search were incorporated without a glance into the copyright owner's §512(c) notification.

The brief also identifies a file entitled "harry potter book report.rtf" whose name and tiny size (1K) make obvious that it is not an illegal copy of the Harry Potter movie. Obvious to anyone who looks, anyway. But why should the record and movie companies bother to look? They're unlikely to suffer any damages if ISPs take down the wrong files, and the consumers involved are unlikely to sue them. (In filing with the Internet Service Providers, a company representative even certified in writing "that we have a good faith belief that use of the material ... is not authorized by Warner Bros. ... or the law." Puhleez.)

Much like the operators of rigged traffic cameras, they're relying on their own institutional power -- and the hassle of opposing them -- to let them get away with near-criminal sloppiness. It's bad enough that you might lose your Internet connection because of such carelessness -- but you could wind up in even worse trouble.

Though no one has gone after the record company whose software seems to have "admitted" that it holds the copyright to the kiddie porn images mentioned above, the FBI is expanding its investigation of kiddie porn. This is a field where, as Wired magazine recently noted, the rule is already "one click and you're guilty," even if you don't realize that what you're downloading is kiddie porn. The FBI will probably start using software 'bots to look for violators. So far, the authorities' record at getting the technology right is unimpressive, as Wired reports:

At the bail hearing for Johnston, Tinney and three other defendants in Houston, the FBI's Kristen Sheldon ... testified that an IP address is, "in very simple terms, a Social Security number. Only one person at one specific time can have that number." In fact, an IP address identifies a computer, rather than a person, and may not even consistently map to a particular machine in networks that use dynamic IP addressing. Midway through the hearing, the presiding U.S. magistrate asked, "What are GIF files?" (Emphasis added.)

This combination of cluelessness and irresponsibility is, unfortunately, not unusual. It also isn't challenged enough. As David Carr writes, who's going to stand up and complain when, if you do, some idiot will probably accuse you of being soft on child pornographers? But those idiots are, well, idiots. When the power to enforce the law is delegated to software employed by people who don't -- or can't be bothered to -- understand it, no one is safe. When you hear that people are using machines to enforce the law, remember the old computer-geek saying: "Garbage In, Garbage Out."

Ask why -- at a time when ordinary people are being asked by politicians and corporations to take more personal responsibility for their actions -- the people who claim to be enforcing the law aren't willing to take responsibility for what they do with their machines. And ours.

Glenn Harlan Reynolds is a law professor at the University of Tennessee and publishes InstaPundit.Com. He is co-author, with Peter W. Morgan, of The Appearance of Impropriety: How the Ethics Wars Have Undermined American Government, Business, and Society (The Free Press, 1997).

Respond to the Writer