Computer

Decoding Speech: The Implications of Granting First Amendment Protections to Computer Code

Decoding Speech: The Implications of Granting First Amendment Protections to Computer Code

Today in the United States, anyone equipped with the correct string of 0s and 1s can 3D print an unregistered, untraceable assault rifle. That’s thanks to Cody Wilson, a self-proclaimed anarchist who struck a deal in 2018 with the US State Department that allowed him to share downloadable, open-source code for 3D-printed weapons. While this may seem like an archetypal Second Amendment problem, the printing of such weapons has instead been made possible by recent interpretations of the First Amendment. It is clear that guns are not qualified as speech. Instead, the settlement was reached on the grounds that banning the release of this computer code violated Wilson’s First Amendment rights.

The code makes the world go round: it lets people communicate with friends, access digital bank statements, send Venmo requests, watch movies, and more. The code also directs critical infrastructure, from flight control systems to electronic medical records to entire power grids. Although unintelligible to most Americans, computer code is its own language with many unique dialects. Code looks and acts like speech, but whether programming languages ​​are protected by the First Amendment is another matter altogether.

The question may seem trivial, but the ramifications of granting First Amendment protections to computer code would be immense. If classified as free speech, it would be virtually impossible to regulate the internet, and the consequences could even extend to public safety. Despite the appeal of the argument that computer code — like any other form of writing — constitutes freedom of speech under the First Amendment, courts must be careful not to legitimize code that is completely unregulated. This would open the door to dangers ranging from a proliferation of 3D-printed weapons to the deregulation of content available on the internet.

“Today in the United States, anyone with the right string of 0s and 1s can 3D print an unregistered, untraceable assault rifle.”

Before assessing whether code should be protected speech under the First Amendment, it is essential to first define what constitutes speech. The Supreme Court established two types of speech: “pure speech,” which is written and verbal, and “symbolic speech,” which includes expressive actions such as choice of clothing, desecration of the flag, and burning of the cross. When lower courts disagreed on the regulation of speech, the Supreme Court upheld that governments cannot restrict the content or message of speech, with some exceptions. According to the Court, it is not the role of the state to determine what the public should see or hear. On the contrary, the duty of government is to ensure that all speech can be heard, whether it offends or not.

When actions affect the public good interest, however, governments have the ability to infringe most constitutionally guaranteed rights. Speech is no exception: governments can adopt regulations with side effects on speech as long as those regulations are content-neutral. Cities have noise ordinances that prevent individuals from shouting after certain hours, political demonstrations cannot take place without permission, and defamation is illegal.

Whether or not First Amendment protections should be imputed to computer code poses an increasingly pressing and unresolved legal dilemma. The issue was first considered by a United States Court of Appeals in Bernstein v. US Department of Justice (1996). The case examined whether the government could stop UC Berkeley graduate student Daniel Bernstein from exporting his code for “Snuffle,” an end-to-end encryption algorithm he created. In an unprecedented decision, the Court ruled that Bernstein’s code could be exported without regulation because his code was protected speech under the First Amendment. The opinion stated: “This tribunal cannot find any significant difference between computer language…and German or French…computer language is just that, a language, and it communicates information either to a computer, or to those who can read it.”

The precedent created in Bernstein was further developed in Universal City Studios, Inc. vs. Corley (2001), when a US District Court ruled that communication does not lose the protections of speech simply because it is written in computer code. The ability of a large audience to understand the speech has nothing to do with the constitutional inquiry. By the turn of the century, there was no longer any doubt: computer code is speech, at least according to American courts. But what about the product of this code? Indeed, it is much more difficult to know if the outputs of the code (websites, video games, algorithms, etc.) are considered protected speech or not.

“The question may seem trivial, but the ramifications of granting First Amendment protections to computer code would be immense.”

Theoretically, the code constitutes pure speech while its outputs could be considered as expressive speech. Although the distinction between First Amendment protections of code and computer programs may seem like a tedious legal question, it has recently come up in relation to both the manufacture of firearms and terrorist attacks, signaling its importance in the real world.

The 2015 San Bernardino shootings brought the issue of code and its outcomes as protected speech to the forefront of American legal scholarship. In an attempt to access the deceased shooter’s communications, the FBI issued an order forcing Apple to create a “backdoor” to the iPhone that would secretly bypass the phone’s security encryption system. The FBI also sought Apple’s cooperation in designing a software update that would remove a security feature that deleted phone data after 10 failed password attempts. The government claimed law enforcement had the legal right to obtain communications through a warrant, but often lacked the technical ability to do so. Therefore, companies like Apple should provide user data to the government upon request. Apple, on the other hand, argued that acquiescing to the order would create privacy vulnerabilities that could be exploited by criminals and foreign agents.

Among other allegations, Apple’s lawyers argued that requiring the company to code software to perform functions for the government “amounts to forced speech and discrimination of viewpoint in violation of the first amendment”. Apple has correctly stated that whenever the government seeks to enforce speech, First Amendment protections are triggered under the doctrine of “state action.” However, the Supreme Court allows government action that infringes free speech if that action is the most narrowly tailored means of serving a compelling public interest; there is a compelling argument to be made that in this case it was. In the end, the government managed to gain access to the encrypted communications before the California magistrate issued his decision. The case was then dismissed.

While Apple’s argument that code is speech is tempting, Cody Wilson proves it’s also incredibly dangerous. If the Supreme Court accepts the argument that the code is pure speech, the ruling could render institutions unable to regulate content shared on the Internet, even in the interest of public safety. Despite the precedent of the lower courts establishing the code as freedom of speech under the First Amendment, the Supreme Court must not open the door to the entirely unregulated distribution of the code.