A judge's wrongheaded interpretation of the federal Computer Fraud and Abuse Act illustrates the problems of allowing civil enforcement of a criminal law.
The case, decided last Monday, arose out of a dispute between two competing companies, Inquiry Management Systems (IMS), and Berkshire Information Systems, both of whom tracked magazine advertisements for their clients. Employees of Berkshire obtained a userid and password from a client of IMS, and used them to access IMS's website and tracking service. This act violated the customer's agreement with IMS.
From there, the Berkshire employees either read, or downloaded (or both) certain copyrighted information about the tracking of magazine advertisements, which of course, they used to compete with IMS.
Is this an unfair and deceptive trade practice? Sure! Inducing a breach of contract between IMS and its customer? Absolutely! Fraud? Sure, why not.
But IMS sued Berkshire for computer crime, and a violation of the DMCA.
On the DMCA claim, the court accepted the argument that IMS's data was copyrighted, and that the passwords designed to restrict access to them were a "technological measure" designed to prevent access to a copyrighted work -- the magic incantation of the DMCA. However, the court made a distinction between what the law prohibits -- "circumventing" such a technological measure -- and what it concluded Berkshire had done -- simply avoiding the measure. Or, as the court concluded, what Berkshire "avoided and bypassed was permission to engage and move through the technological measure from the measure's author." Not a DMCA violation.
But the court took a more expansive reading of the federal Computer Fraud and Abuse Act, which punishes anyone who:
intentionally accesses a protected computer without authorization, and as a result of such conduct, recklessly causes damages; or intentionally accesses a protected computer without authorization, and as a result of such conduct, causes damages; and by [this] conduct ... cause[s] ... loss to 1 or more persons during any 1-year period ... aggregating at least $ 5,000 in value ...
The statute defines "damage" as "any impairment to the integrity or availability of data, a system, or information" and allows a civil lawsuit to be filed by "[a]ny person who suffers damage or loss by reason of a violation."
According to Judge Buchwald in the Southern District of New York, Berkshire violated this law. The court reasoned that using the userid and password in violation of a contractual provision was an unauthorized access.
The fallacy of this decision becomes clear if you consider that the customer himself could have logged in with his or her password, obtained the documents and records, and given them to Berkshire -- and this would have been a simple breach of contract, not a crime.
Under this decision, anytime you share your userid and password you're potentially committing or facilitating a felony. If you subscribe to, for example, the New York Times website (a free login) and let someone else log in with your userid, you might end up in the slammer -- at least if you appear before Judge Buchwald.
The real problem with Judge Buchwald's analysis was her willingness to accept IMS's argument that the "access" by Berkshire cased "damage" -- that is, the impairment, modification or potential impairment of the computer system. The court accepted as factual IMS's argument that, by Berksire's copying IMS's data and using it to assist in the creation of its own competing system, Berksire inflicted "irreparable harm" on IMS, "as the integrity of its data and system was impaired."
Certainly IMS suffered legal "damages," from how did the "copying" affect the integrity of the data and system? This is similar to the bogus argument that was initially accepted in the criminal prosecution of Brett "Secret Squirrel" McDanel, after he told people that their e-mail was not secure. The company was "harmed" and "damaged" by the revelation, but this does not mean that any data was impaired or modified.
The terms "damage" and "loss" as used, not only in the computer crime statute, but others as well, are only vaguely defined -- creating substantial ambiguity in the criminal and civil law. For example, a few months ago the St. Paul Insurance company rejected a claim by AOL for indemnification of lawsuits filed by users of AOL 5.0. The users claimed that the software caused their systems to crash, requiring not only rebooting, but reformatting of hard drives. AOL paid a settlement, and sought indemnification by its insurer, since its policy covered losses resulting from "physical damage" it caused to others. The court concluded that wiping the data from the hard drives and rendering them inoperable was not "physical damage" and therefore no recovery from the insurer was possible under the policy.
The case in New York demonstrates several things. First, criminal statutes that allow civil lawsuits (like the RICO statute and others) inevitably result in distortions of the criminal law -- with potentially bad consequences for criminal defendants. They unnecessarily expand the scope of what is criminal.
Second, even though the statute has been amended at least a half a dozen times, it's still hard for the courts to tell the difference between "damage" -- harming the system -- and "damages" -- harming the owner of the system in a way that results in quantifiable loss. It's time for Congress, or the Supreme Court, to clarify the matter once and for all. And while they're at it, they should take another hard look at what's meant by "unauthorized" access.