There is growing momentum in Washington to regulate AI: a recently signed executive order requires review of AI models, congressional proposals aim at further AI legislation, there is discussion of the government potentially taking equity in frontier AI labs, and in a recent Friday action foreign nationals were prohibited from accessing Anthropic’s most advanced models. The authors view these moves as the possible opening salvos of a broader regulatory push.
They warn that future measures could, intentionally or not, regulate or even ban open source—an outcome they consider a serious mistake. By open source they mean the public, transparent process of sharing, building and distributing technology. They argue open source is safe, secure and economically beneficial.
Scale and history
The piece cites that more than 90% of the world’s software has long been built on open source, and that open source produced over 8 trillion dollars in economic benefits well before AI’s rise. Today, open source components quietly support the training, improvement, deployment and securing of AI systems across the economy.
The movement’s roots lie in academia and the free software tradition. The free software movement began in 1983 on the campus of the Massachusetts Institute of Technology (MIT), at a time when software use often meant paying or negotiating with large corporations. Since the rise of open source, students and learners at every level — from universities and community colleges to programming bootcamps — have relied on freely available code and tooling as essential parts of technical education.
Three values: education, innovation, competition
- Education: open source gives students and learners access to real code and tools, enabling practical training in programming and engineering.
- Innovation: open source combines reusable tools and active communities so that ideas can be turned into working projects without prohibitive cost. The authors note that some major companies—citing Meta, where the initial version of Facebook was developed on an open source stack—have roots in this ecosystem.
- Competition: open source levels the playing field, allowing smaller players to challenge dominant incumbents. The article points to Linux as the operating system that now runs more than 90% of the world’s cloud infrastructure and acted as an antidote to Microsoft’s dominance; it also recalls that former Microsoft CEO Steve Ballmer once labeled Linux “cancer.” Android’s open nature similarly fostered long periods of competitive activity in the smartphone market.
According to the authors, these roles remain valid in the AI era. Closed, proprietary models from players such as Anthropic and OpenAI are concentrating power; Anthropic has, they say, reduced its most advanced model’s capability when it is used to improve another party’s model. Open-source AI—especially open-weight models—have served as the only practical counterweight for startups, educational institutions and enterprises seeking alternatives.
Safety, transparency, and the China question
The authors concede that the security implications of open source models reaching frontier capabilities merit monitoring. Still, they argue that the intrinsic transparency of open source tends to make systems safer and more secure, because more engineers and researchers can detect and correct unwanted model behaviors or software bugs—summarized by the adage, “given enough eyeballs, all bugs are shallow.” They also note that an open source model installed on a company’s own infrastructure need not transfer data externally, a point raised by Brian Chesky, CEO of Airbnb.
On geopolitical risk, they caution against using competition with China as a pretext to regulate open source. Chinese-origin open source models are already improving the efficiency of many American startups, especially those that cannot afford the monopoly-level prices charged by Anthropic or OpenAI. Rather than restricting open source because some models originate in China, the authors suggest the U.S. should invest more in open source domestically. Limiting open source would chill education, innovation and competition, and could push other countries to align with Chinese alternatives.
Conclusion: open source as sunlight
Invoking former Supreme Court Justice Louis Brandeis’s remark that “sunlight is said to be the best of disinfectants,” the authors portray open source as this kind of sunlight in technology and AI. Their recommended stance is for the United States to favor transparency and openness, preserving the educational, innovative and competitive benefits that open source delivers, rather than treating it as a regulatory target.
They close by asking readers to share the op-ed with anyone undecided about open source AI or new to the topic.
