An article in U.S. News & World Report.

By Vilas S. Dhar – Contributor

Advances from DeepSeek and Alibaba show we can democratize AI with faster models that are cheaper to produce and easier to use.

Mark your calendars: This is the week that conventional wisdom about artificial intelligence was turned on its head, and with it, all of our assumptions about the future of AI.

That’s because a small Chinese startup named DeepSeek accomplished what many thought impossible: building an AI system that rivals ChatGPT’s capabilities at a fraction of the cost and making it freely available. DeepSeek’s free mobile app swiftly dethroned OpenAI’s ChatGPT as the most-downloaded free app in the U.S. on Apple’s App Store. Days later, the Chinese multinational technology company Alibaba announced its own system, Qwen 2.5-Max, which it said outperforms DeepSeek-V3 and other existing AI models on key benchmarks. What we’re witnessing is unprecedented: the democratization of artificial intelligence beyond the control of any single nation or company.

For years, building sophisticated AI required massive resources that only Silicon Valley tech giants could muster. Just last week, San Francisco-based OpenAI announced the Stargate Project, a new venture backed by SoftBank, OpenAI, Oracle and MGX to the tune of $500 billion – an astronomical amount to spend on next-generation systems. Alongside the billions of dollars that Google and Microsoft have poured into AI infrastructure, the momentum to invest more and more seemed to confirm that bigger spending equals better AI.

Now DeepSeek, Alibaba and others have disproved that formula. We’re seeing a proliferation of AI capacity – faster models that are cheaper to produce and easier to use – emerging from China, India, Europe and other markets.

When AI development once required billions of dollars, only the largest companies could participate. Now, that barrier is crumbling – and with it, our assumptions about who can lead in artificial intelligence.

Silicon Valley built its technological leadership on private sector dominance, where proprietary data and concentrated resources drove innovation. The United States reinforced this approach by restricting access to advanced computing chips, believing this would maintain our technological edge.

But these restrictions have instead accelerated innovation elsewhere, spurring investments in alternative approaches and new chip designs. The rapid democratization of AI capabilities demands a different strategy.

The growing movement toward low-cost and broadly available AI fundamentally challenges how innovation spreads. When sophisticated AI systems become accessible to a broader community of developers and researchers, they can be adapted to serve local needs for health care, education and a host of additional industries.

As a member of the United Nations’ High-Level Advisory Body on AI, I’ve worked with global experts to envision frameworks for a more cooperative technological future. Our conclusions in the September 2024 UN report “Governing AI for Humanity” identify the critical need to expand public access to data, create targeted funding mechanisms and to build technological capacity across communities – enabling local hubs of innovation that bring fresh perspectives to these powerful tools.

Local hospitals can develop health care tools that understand their patient populations. Rural schools can build learning systems adapted to their students. Small businesses can create AI solutions tailored to their specific markets. Communities can build AI models to address local impacts of climate change – and share these products at low cost with others across the globe. And these are just the tip of the innovation iceberg.

America stands well positioned to lead this new era, but seizing this opportunity requires us to update our playbook.

Critics will rightly point out the risks of democratizing such powerful technology. More accessible AI could make it easier for bad actors to create harmful applications, from sophisticated cyberattacks to targeted disinformation campaigns. Legitimate concerns exist about quality control and safety standards when AI development moves beyond the walls of well-resourced labs. And U.S. tech companies, which have invested billions in proprietary AI systems, will face real economic challenges in an open-source world.

This vision of a distributed AI future must also contend with obstacles: Running advanced AI still requires expensive computing power and data centers. But new solutions are emerging – from shared computing networks to more efficient AI models that run on smaller computers. The hardware barrier, like many before it, is already starting to splinter.

These are serious concerns that demand thoughtful solutions. But they actually strengthen the case for American leadership in shaping an open AI ecosystem. Rather than restricting access to advanced AI chips and keeping our technological cards close to our chest, we can address a broader array of risks by collaboratively creating frameworks for responsible innovation – including security standards, safety testing protocols and clear liability rules.

Addressing these possible risks also requires us to invest in AI literacy as comprehensively as we do in reading and writing, building security frameworks for an open-source world and creating public infrastructure that helps communities participate in AI development safely. Just as we did with the internet, America can lead in developing the governance structures that make technological openness work for everyone.

America’s greatest technological achievements have always come from creating environments where innovation flourishes freely. By embracing open innovation while promoting our democratic values, we can ensure these powerful tools evolve to benefit everyone.

The future of AI will be distributed, collaborative and open. America’s next chapter of technological leadership depends not on controlling who gets to innovate, but on creating the world’s most powerful ecosystem for breakthrough ideas.

That’s a future worth building.

Vilas Dhar is a global AI policy expert and president of the Patrick J. McGovern Foundation, a philanthropy focused on exploration, enhancement and development of AI and data science for the common good.