← Back to Posts

A hacking law written after WarGames: Daniel Takash on the CFAA

Sean Roberts•December 10, 2020•3 min read

I opened my conversation with Daniel Takash, regulatory policy fellow at the Niskanen Center, with Aaron Swartz. Swartz downloaded a large number of academic articles from JSTOR over MIT's network, and federal prosecutors charged him under the Computer Fraud and Abuse Act. Daniel was careful to say he is not a criminal justice expert and would not second-guess the Justice Department or the MIT police. His point was about the statute. Swartz faced a possible thirty-five years in prison for conduct that, by Daniel's account, a JSTOR representative called inconsiderate but did not think deserved criminal penalties. He died in 2013, before any trial.

If a law can produce such tragic consequences, no matter the intentions, I think that should give everyone pause.

Then he told the origin story, which I had heard before and still enjoy. As he tells it, members of Congress in the 1980s watched WarGames, in which Matthew Broderick's teenager dials into a military computer and nearly starts a nuclear war, and noticed there were no criminal penalties for what they had just seen. Maybe not the greatest start, Daniel said, to a serious portion of the criminal code.

The trouble is in the wording. The statute punishes accessing a computer without authorization, which almost nobody objects to. It also punishes exceeding authorized access, which nobody can define. Representative Zoe Lofgren and Senator Ron Wyden tried to clarify that phrase after Swartz died, in a bill people called Aaron's Law, and it went nowhere.

Daniel's illustration came from a law professor. Your company forbids personal financial business on work computers. At your desk you remember that your credit card bill is due, so you log in to your bank and pay it. On a plain reading of the statute you have exceeded your authorized access. Nobody thinks you will go to prison, and your company may not care. But the law allows it, which means it can be pointed at anyone whose behavior on a computer somebody does not like.

Which brings us to Van Buren v. United States, argued at the Supreme Court on 30 November. The facts are not sympathetic. A police officer used a law enforcement database he was entitled to use, for a purpose he was not, and was convicted under the CFAA. The question for the Court is whether misusing access you legitimately have is the same crime as breaking in. Daniel worried that bad facts could make bad law, but he was relatively optimistic that the justices will read the phrase narrowly.

From the engineering side, the narrow reading is the only one that makes sense. Most people who write software have broken an acceptable use policy at some point: personal mail on a work laptop, a script pointed at an internal system nobody formally approved. If those are federal crimes, the statute is not a hacking law. It is a way to turn an employer's handbook, or a website's terms, into a prosecution whenever someone wants one. A law built for breaking in should stay a law about breaking in.

We should know by the end of the Court's term next June. Daniel's bigger worry, though, was not individual users. It was what the same phrase lets a platform do to a competitor, and I have written that half of the conversation up separately: terms of service with a federal backstop.

From Lincoln Shorts, the interview series I hosted for Lincoln Network, recorded 10 December 2020.

Share this post