The First Amendment to the US Constitution, a pivotal pillar of American democracy, resolutely proclaims: “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.” This amendment, especially the section about freedom of speech, is deeply intertwined with the current discourse surrounding the US drive to ban open-source AI.
AI isn’t the first instance where the US has attempted to ban software’s free distribution and usage. During the Cold War, it was illegal for companies to export robust encryption algorithms, including their source code, from the US. However, the courts determined that the source code is speech protected under the First Amendment. Therefore, let’s delve into the history of open-source as a form of speech and how that might shape the ongoing debate around AI, underscoring its significance.
The History of Encryption and the Freedom of Speech
Since antiquity, states and merchants have used encryption to obfuscate and secure communication and information. Yet, in many cases, the communication only had to be secured sufficiently long enough to make it useless. For example, most militaries didn’t care about protecting orders after the campaign. Additionally, states could circumvent most encryption by simply adding more personnel. Technological advancements, such as the Enigma and the responding Colossus, changed the calculation. Relatively simple machines and algorithms require ever-increasing power to break.
The US responded by banning the export of encryption as ammunition and actively investigated anyone trying to export the source code. Consequently, the source code for GPG was exported as a book and printed on T-shirts, as both actions enjoined solid protection under the free speech law.
Ultimately, US courts decided that source code is a form of speech, whether printed in a book or distributed digitally. Subsequent executive actions have tried to tweak the rules but haven’t been tested again in court.
The Drive to Ban Open-Source AI
The drive to ban open-source AI to prevent US adversaries from gaining access to the technology is similar to the restrictions on cryptography. From disinformation via privacy concerns to personalized ransomware, there are many ways that nation-states and criminals can exploit AI. Each is an excellent target for politicians to show them doing something.
Yet, in all these cases, regulation will help the most powerful defend their vision of the future and their bank accounts from the disruption AI might bring. It restricts access to AI for a wide range of society and ensures that big tech can continue building out its monopolies.
Banning open-source AI will prevent innovative minds from building ever-newer models and challenging incumbents who are too slow to act or concerned about hidden agendas.
Worse, it will do nothing to prevent criminals and foreign governments from using AI. Foreign governments and cyber criminals either have the means to develop their own AI or can break into corporate systems to copy proprietary AI. Countries like Russia have exceptions in their criminal and intellectual property laws that legalize software piracy, even by espionage, if it is in the national security interest.
Thus, law-abiding citizens and societies will only be harmed if open-source AI is banned. Criminals and advisories won’t care.
Use your Freedom and Speak!
While the courts might give a backstop to any outright ban of open-source AI, litigation is costly and time-consuming. Supporting organizations like the EFF can help share the burden. Likewise, writing your representatives and lawmakers can help sway their minds.
However, the most crucial step for us is to speak out about the success of open-source software and continue to create open-source AI. To have your speech protected, the first thing you have to do is speak!