Doomsaying is a terrible basis for public policy. Still, some of America’s top AI leaders are deploying that exact sort of rhetoric. Anthropic’s Dario Amodei has repeatedly warned about the dangers of open-source AI, anointing himself as our generation’s cautionary priest on AI, despite leading a leading frontier lab seeking an initial public offering valuation near $1 trillion. Analysts are calling this “doom trolling.”
If Amodei has his way, open-source models would effectively be sent to software jail, never to be seen again. While he is entitled to his concerns, Amodei’s conclusions are cynical and backward. The United States must not label open-source technology the enemy. It’s key to winning the tech race, which will define the 21st century.
Recommended Stories
America is the envy of the world in technology for a simple reason: It didn’t hide from the future. It set out to build faster and better than anyone else, and did. As Neil Chilson put it when testifying before Congress in June, “We will not secure another quarter millennium of technological leadership by abandoning the principles that have brought us this far.” Those include competition, openness, diversification, and due process — in other words, the opposite of Europe’s approach to technology: safety first.
A trip down memory lane should remind us why limiting the digital ecosystem to proprietary systems would be a mistake.
The internet flourished in no small part because TCP/IP wasn’t proprietary. Linux went from a volunteer software project to an entire operating system helping power parts of the world’s digital infrastructure.
AI can offer the same opportunity, if we let it.
While today’s frontier models may emerge from multibillion-dollar frontier labs, the breakthroughs won’t all come from just a handful of proprietary ecosystems. They’ll come from universities, startups, and communities where countless people are constantly contributing and iterating. Every experimentation an engineer pursues by leveraging open-weight models acts as a force multiplier, driving American innovation toward use cases no one could have imagined.
This open-source ethos of challenging the status quo sparks the dynamism that makes America so special. It’s also why efforts to clamp down on open-source AI are so misguided.
On the world stage, America’s chief competitor, China, has no intention of shutting down open-source progress. Beijing correctly recognizes that open-source AI is a tool for exerting its political influence on the world stage. Its companies are aggressively releasing increasingly capable open-source AI models that anyone around the world can download and deploy.
It is a complement to the broader Digital Silk Road strategy: export digital infrastructure and platforms that deepen the Chinese Communist Party’s influence. It has made inroads in Asia, Africa, and Latin America, and it has its sights set on expanding this soft-power influence.
Should international firms and governments both begin to rely on Chinese models, the soft forms of U.S. dominance we’ve benefited from for decades will vanish.
While Amodei and others may argue that restricting open-source models or implementing export controls are necessary to keep advanced AI out of hostile hands, such measures are wishful thinking. Determined adversaries will continue to develop their own models, acquire open models from elsewhere, or build on existing work.
Consumers, startups, small businesses, and our allies ultimately stand to lose when restrictive policies prevail. They look to America for technological leadership. We make leading that much more difficult when we ask our own innovators to compete with ankle weights while our rivals sprint to close the gap.
Restricting open-source AI would also narrow the field of innovation to only a handful of well-resourced firms, forcing many talented and creative engineers to the sidelines. The end result would be less competition, fewer builders, and fewer opportunities for the kind of rapid innovation and iteration from open-source software that helps drive the country’s edge.
AMERICA BETS ANTI-CHINA AI STRATEGY ON GOVERNMENT THAT FUNDED 421 FAKE INFRASTRUCTURE PROJECTS
And while, yes, frontier AI models do raise novel questions and pose broad risks, particularly for cybersecurity, there is a huge difference between mitigating specific risks and embracing a philosophy oriented toward technological scarcity.
Amodei is wrong to seek a clampdown on open-source tech. America has historically won because we have a dynamic environment that rewards breakthroughs, mold-breaking, and competition, including open-source technology. We cannot surrender that edge just because a doomer-pilled CEO asks us to.
James Czerniawski is the head of emerging technology policy at the Consumer Choice Center. His work has been featured in the New York Post, Newsmax, Newsweek, the Daily Wire, and more. Follow him on X @JamesCz19
